Transcript of Episode 89: Beth Cimini

Disclaimer: Transcripts are automated and may contain errors.

Grant Belgard: Welcome to The Bioinformatics CRO Podcast. Today, I’m delighted to welcome Dr. Beth Cimini, who leads the Cimini Lab within the Imaging latform at the Broad Institute of MIT and Harvard. Beth works at the intersection of microscopy, computational biology, open-source software, and scientific community building. Her group helps researchers turn images of cells into quantitative reproducible biological measurements and develops open-source bioimage analysis tools, including CellProfiler, Piximi, and BiLayers. Beth’s career has spanned biochemistry, molecular biology, high-content imaging, and software, with a path from Boston University to a PhD at UCSF and then to the Broad.

Grant Belgard: We’ll talk about what bioimage analysis can teach us about biology, how scientific software and communities actually get built, and what advice she has for people who want to work across experimental and computational biology. Beth, welcome to the show.

Beth Cimini: Oh, thank you so much for having me. I’m delighted to be here.

Grant Belgard: For listeners who are new to bioimage analysis, what’s the core problem you spend your time trying to solve?

Beth Cimini: Yeah, so mostly we deal with images that come from light microscopes, and it feels like light microscopy should be a solved problem because we invented it in the 1600s, right? But it turns out that we invented light microscopes about 300 years before we invented computers, and so there’s a fantastic diversity of biology that we can study under the microscope. And cells might be anywhere from a pixel to a whole field of view. There’s tons of different fluorescent antibodies and can have many channels. And when you’re a bioimage analyst, you’re presented with an image and you’re sort of told, find the interesting biology here.

Beth Cimini: And so there’s a couple different pain points, one of which just being how can you find the objects of interest in the first place, which feels like it should be simple, but some of that’s because we have very good onboard neural networks, let’s call them, that help us find the boundaries of things, because otherwise we’d run into stuff all the time. Computer neural networks are still getting as good at object detection and object recognition as our brains have been, just because we’ve been doing that for a lot less time. And then there’s all sorts of sort of subtleties and nuances around, you know, how when you find the objects, can you measure them and interpret what interesting biology is happening. So it’s sort of a million dimensional problem, but it’s what makes bioimage analysis a lot of fun and very rewarding.

Grant Belgard: What does a typical week look like for you?

Beth Cimini: Oh, gosh. Well, so now I’m a PI, so now a typical week looks like a lot of emails and grants. But our lab is really fun. We’re sort of a four-legged table in that we do a mixture of open source software creation for image analysis, so things like CellProfiler and Piximi and BiLayers that you mentioned. We do image analysis methods research. So how can we make some of these things better and faster? We do image analysis almost as a CRO. So groups come to us, we collaborate with them from academia, pharma, biotech, wherever. And we do image analysis education and outreach. So a typical week, we’re touching all of those bits a little bit. And so we have an online educational platform that we’re building.

Beth Cimini: So it might be sort of checking out lessons that we’re building there, designing what the lessons are that we need to sort of put to make it easier for people to learn bioimage analysis, might be sort of checking in on a collaboration and seeing, you know, how’s it going for the workflow for, you know, these cool 3D images that, you know, our collaborator just sent over, and then sort of fixing software bugs and a little bit of everything. It’s chaotic, but it means it’s never boring.

Grant Belgard: What kinds of scientific questions tend to bring people to bioimage analysis?

Beth Cimini: Yeah, light microscopy is still one of the best things and certainly the most cost effective thing we have for doing anything where you want to do single cells. We’ve been looking at single cells under microscopes, again, for hundreds of years. So and especially anything where you want to do biology over time, we really can only of the omics that are out there only sort of reliably do sort of omics over time in light microscopy, because it’s the only omic we can do while the cells are still alive and still happy and still mostly doing their normal biology. But really, you know, we’ve estimated from some very back of the envelope things that we think probably about a third of biologists do some sort of light microscopy. So it can really be almost anything. Because we’re at the Broad and the Broad’s mission is to do things at scale.

Beth Cimini: We tend to work with people who are doing like early stage drug discovery, high content screens, but that’s a function of where we are. Biomage analysis definitely happens everywhere. It makes microscopy images hard to turn into trustworthy measurements. I think the thing that is tends to be the sort of most surprising to folks who don’t spend a lot of time thinking about it is, you know, for a lot of other modalities, we can very cleanly separate, you know, when we’re doing a sequencing read, and we see, you know, that we have a change, we have a nucleotide polymorphism that is sort of causing a biological change. Generally speaking, the change in the sequence doesn’t change our ability to do sequencing. Most biological changes won’t affect the actual measurement.

Beth Cimini: But when we get to bioimage analysis, we’re trying to look directly at the, you know, cell object or nuclear object or worm or whatever sort of object you might care about. And the phenotypic changes that are usually exactly what we want to detect, are going to affect our ability to find that object, potentially, you know, if we have a very finely tuned, you know, object detector or segmenter, to find, you know, cells, and we’ve told it, you know, cells are between 10 and 15 microns in diameter, if we have cool biology that causes the cells to be 17, our algorithm might just stop detecting the cell altogether. And so we will miss that cool biology, because it has gone outside the, you know, the definition of what a cell is.

Beth Cimini: And so it’s very hard to cleanly separate what’s a quality issue, and what’s an actual interesting biology, because we’re doing sort of detection of the things we care about, and sort of phenotype of the things we care about all in the same step. It’s not that we’re like, doing sequencing where we can measure if we’re getting a good read, and then sort of making inferences, you know, once we’ve collected our read, it’s the very things that we’re trying to detect as our biological, initial biological measurement, or the biology, there’s no intermediate steps. And so how can we do this well? And how can we make sure that we’re finding all the variability we want, and not other things like a, you know, 18 micron piece of crud that’s sitting there can be really challenging.

Grant Belgard: On that note, what makes a collaboration between a biologist and an image analyst successful?

Beth Cimini: That’s a great question. It’s one we spent a lot of time thinking about. It really, I think, just comes down to communication. We run office hours just for helping folks sort of, you know, get over particular small problems and things like that. But there’s a lot of nuances around, depending on exactly what kind of measurement you want to make. So example, if you’re taking measurements of colocalization, which are one of the most popular things to want to measure, you know, are two molecules in the same place, you know, maybe that means they’re interacting, maybe it doesn’t. There are certain kinds of controls that are really critical to do and will change the kinds of measurements that you can sort of reliably interpret if you’ve made them or not.

Beth Cimini: Which kinds of measurements are appropriate for which kinds of problems is sort of a thing that is like a passed down legacy from sort of bioimage analyst to bioimage analyst. But it’s not really anything where there’s hard and fast rules. There’s a lot of, you know, try this and, you know, under these circumstances, you know, this works and under the other circumstances. So it’s just a tremendous amount of, like, integrated knowledge that somebody who’s used to thinking about their biology as objects and not about how we turn those objects into quantitative measurements has just never contemplated. And so just lots of conversations. The good thing is all of the bioimage analysts that I know, and I know a lot of them are all, like, really, I describe us as a community of kind nerds who want to help people.

Beth Cimini: It’s really a field that tends to attract people who love collaborating and love talking about the sorts of stuff and sharing this knowledge and helping put people on the right path. It’s honestly the, like, the friendliest community I’ve ever been a part of. It’s really lovely.

Grant Belgard: How do you explain the difference between making beautiful images and extracting useful measurements?

Beth Cimini: I mean, they can be the same thing. They certainly can be. You know, a clean image that is mostly full of debris and stuff, you know, can’t, is probably going to be beautiful. Of course, beauty is in the eye of the beholder. But as a human, a lot of our, you know, visual system, we’re drawn to contrast. And so sort of people will do things like play with the gamma, which is, I’m doing hand gestures that, of course, the audience can’t see right now. Essentially, how much we change the value of the pixel that you see based on sort of the brightness of the photons that are hitting the detector, you know, because for human eyes, we like to see contrast, you might make it so that really small changes in pixel intensity sort of look very different and very striking and beautiful in an image.

Beth Cimini: But of course, we don’t want to artificially inflate the differences between two parts of our image when we’re trying to make detailed quantitative measurements. And so the things that will help you make quantitative measurements will also help make your images prettier in that, you know, they’re in focus, they’re clean of debris. But there’s a lot more we’re allowed to do if we’re just going for a pretty picture.

Grant Belgard: What kinds of projects are most fun?

Beth Cimini: Oh, gosh. I really like personally working with folks who are just getting into this space. So, you know, maybe this is the first time they’re sitting down and doing bioimage analysis, because just the like, the moment when it clicks and you see somebody be like, oh, I get it. I know what to do now. Like, that’s such a fun moment for me as a professional to sort of get to help somebody else get onto the path. I really enjoy that. But we’ve also been part of some some huge, you know, collaborative things like the JUMP-Cell Painting Consortium, which was organized by Anne Carpenter, also here at the Broad, where we were doing analysis for data made at 14 different sites.

Beth Cimini: And that was crazy to try to coordinate everything, but was also really cool to see, you know, it came out with sort of 200 terabytes of images and, you know, the biggest publicly available open source cell painting data set. And so things like that can be an awful lot of fun, too.

Grant Belgard: What kinds of projects are most dangerous to kind of, you know, be analyzed badly.

Beth Cimini: Oh, geez. That’s a good question. I think it’s really easy to do uncareful analysis, like, there’s a lot of big problems with bioimage analysis. There’s lots of hidden mine fields. I would say colocalization, I mentioned earlier, is one of the hardest ones. And that’s because there can be really subtle things with stuff like bleed through with stuff like, if I want to measure how often a little thing is inside a big thing, I need to take into account what fraction of the image is covered by little things and big things. So, you know, I have large spots and small spots. I want to know if the small spots are in the big spots, but I need to know if the whole image is big spots. They might be, but it might be by accident. It might be coincidental. And so colocalization is one of the things people most often want to do, but it’s one of the ones with the most traps in it.

Beth Cimini: So actually for this online educational resource that we’re doing, we started with saying we’re going to do a chapter about colocalization. And now we’re doing a whole course on colocalization because we know it’s really hard and we know there’s a lot of sort of subtleties to get right.

Grant Belgard: What makes cell painting useful for learning about cell state?

Beth Cimini: Yeah. So I can explain cell painting a little bit first for folks who’ve never come across it before. Cell painting is an assay developed here at the Broad in the early 2010s where the idea was, and it was, you know, more of a controversial idea at the time than it certainly is now. We have these cell images. They’re beautiful. They’re full of, you know, particular stains and particular markers to learn particular biology. But we certainly now understand from things like single cell sequencing and other just sort of computational domains, you know, just having high dimensional information, just lots and lots and lots of data means even if we don’t understand any particular specific data point, you know, we can draw connections between sort of samples or things like that. And so this is the way a lot of deep learning works.

Beth Cimini: We don’t really understand what a particular neuron in a particular deep learning architecture does, but we know when the whole thing comes together, you know, something interesting pops out. And so the idea was let’s put as many dyes as we can that will be inexpensive and fit on a standard microscope so that we can create really rich feature descriptions of cells. We’re not going to worry too much about what any individual feature means. And in fact, now in the era of deep learning, we sometimes use deep learning features that nobody has any idea what they mean. And just by doing lots and lots and lots of measurements on lots and lots and lots of cells will create big data big enough such that sort of interesting biology will fall out.

Beth Cimini: Now, I have to say, when I first came to the board, like I was trained as a sort of very standard cell molecular biologist and I was told about this assay and they were working on like the first one of the first big results papers from it. I was very skeptical. I kind of was like, there’s no way that works. Like, why did I just spend the last, you know, eight years of my PhD, like putting fluorescent proteins on particular things if all I had to do was dump a bunch of dyes on there and just sort of like study what pops out. But it turns out that, you know, making these dense feature representations is incredibly powerful. And so in that the eLife 2017 Rohban et al.

Beth Cimini: paper, which was one of the first big results papers from cell painting, sort of showing that you could detect novel pathway-pathway interactions between two different sort of gene pathways that were never previously hypothesized to interact just because they had a connection in cell painting feature space. I was like, whoa. And so now cell painting is a super well-adopted tech assay in the sort of early drug discovery space. It’s fun because it’s really sort of inexpensive to do. At scale, it’s less than $2 a sample, you know, compared to many thousands for a lot of other omics assays. It can be scaled up really easily. It’s amenable to every kind of mammalian cell we’ve ever thrown at it. And I know some folks who are working in non-mammalian cells who are using variants on it too. And you can get a lot of interesting biology out there. Now, it doesn’t work for every phenotype.

Beth Cimini: You can’t detect every kind of phenotype or every type of biology in it. But if you’re lucky enough that you can detect it in cell painting, then you have this sort of really great, really powerful assay that’s super easy to spin up and super easy to scale up if you decide you want to test it on hundreds or thousands of sort of drugs or genetic variants.

Grant Belgard: What kinds of biological variation are images especially good at capturing?

Beth Cimini: Yeah, I think there’s a lot that images are good at capturing. But, you know, a lot of stuff, because it has been the easiest modality to scale first, you know, a lot of what we know about biology comes from sequencing. And I work at The Broad, which is a place that grew out of the Human Genome Project. Sequencing is incredibly powerful and incredibly useful. But we know that there are things like, you know, regulation of translation. There’s an interesting paper that came out just a couple of weeks ago, Nature or Science, I forget which, about alternate translation. Alternate translation is actually a super common thing. Cell behaviors that happen on a fast state that happen faster than transcription or translation. You know, you phosphorylate a molecule, maybe it changes its behavior, maybe it goes somewhere else, maybe it gets sort of kinase gets turned on or off.

Beth Cimini: You can detect that with microscopy, you know, within seconds of it happening. For other omics, you have to wait for sort of like whatever the downstream activity is to sort of build up whether that’s, you know, you’re doing proteomics, and you’re sort of detecting, you know, new additions of ubiquitin or things like that. Those things take time. But for microscopy, you can do things sort of instantaneously, and you can capture cell-cell interactions, and you can capture cell behaviors in ways that are still really hard with other omics. The other omics are absolutely catching up, and all of the omics sort of working together is really our sort of vision of eventually where biology gets to.

Beth Cimini: But we are getting phenotype with microscopy at the level of like, okay, we can see how all the RNAs, all the proteins, all of the other things we mentioned in the other omics come together and make a cell behave.

Grant Belgard: What kinds of biological variation are images bad at capturing?

Beth Cimini: Yeah. So changes in sequence, obviously, are possibly detectable if they cause a downstream phenotype, but maybe not. I mean, it has to, for any sort of perturbation, we need to either have it change the cell enough that we can see a difference in something like a bright field image or a cell painting stain where we don’t know what specific changes we’re looking for, or we need to have a readout or a, you know, sort of something that’s designed to detect a sensor for a particular change. The good news is there are, you know, tons of sensors, and there’s a million antibodies in the RRID catalog, literally, but we need to know which one we want to detect. And unless you’re doing something like imaging mass cytometry or spectral microscopy, you can usually only detect four or five molecules at a time.

Beth Cimini: So if you want to see how 50 things are changing at once, that’s incredibly easy with sequencing. That’s incredibly difficult with microscopy. So parallelization is still, I would sort of sum up where I got to with that.

Grant Belgard: Where can image analysis create false confidence?

Beth Cimini: Oh, there’s a great paper on this recently from folks at Janelia called Believing is Seeing. So people assume because the computer did it, it is therefore unbiased. And having a computer analyze your images is definitely more unbiased than you sort of like doing what I was trained to do when I was sort of still a very young scientist, which was like, just like group all the pictures by like, by what they’re supposed to be pictures of, and then like decide what you think is the difference and then pick out some representative pictures and say representative image shown. That is the least, that is the least sort of unbiased you can be because a human is making those decisions. But again, things like I said about where you are making decisions about what the algorithms find.

Beth Cimini: It can be also really pernicious things like, well, I expected how, when do I decide my workflow isn’t working and decide to change it? Do I always do it no matter what? Or like, oh, in sort of run one of this assay and run two of this assay, I detected a threefold difference. And then the third time I run it, now I don’t see that difference anymore. Now maybe I decide to check, well, maybe my image analysis isn’t working anymore, but maybe it wasn’t working between one and two. And we just never checked it because the results were the same. And we were like, oh, it’s right. And so we only, you know, inspect really carefully when our results are unexpected. And so I definitely recommend reading that paper, seeing, believing is seeing, because if you believe it, you might find it. It’s a good way of how even with quantitative image analysis, you can trick yourself.

Beth Cimini: So it’s absolutely better to do quantification, but we’re humans, we’re biased. And, you know, we just have to, we have to design that into the systems that we build. And that’s something that we do try and design into the systems that we build to sort of make it easier to tell when the analysis is going wrong, but not in a way where we know the results and we expect what the results will be, but just in a more unbiased fashion.

Grant Belgard: What do you think about batch effects and imaging experiments?

Beth Cimini: Oh, batch effects and imaging experiments are like the thing that we are fighting against the most often, you know, especially with cell painting. The wonderful thing about cell painting is it is exquisitely sensitive. The horrible thing about cell painting is it is exquisitely sensitive. We’ve, you know, talking to many folks who are experts in this technique from all over the world over the years, you can tell with cell painting, which plates were in the back of the incubator versus the front of the incubator, which ones are on the incubator shelf versus like are sitting on top of another plate. And so you have all of these really exquisite differences that don’t matter, that are not what you want to detect, but they’re all mixed in with the differences you do want to detect, which are, again, the sort of things in biology.

Beth Cimini: And so being really careful with experimental design when you’re doing an imaging assay, especially something like cell painting, where you’re not always putting your controls in the same places or, you know, you’re making sure maybe you don’t even know what’s plated in each well, like you ask your friend to plate it for you and write it down. And you only sort of like you treat everything in a completely blinded fashion. It’s still really hard. And that’s why we do sometimes need to change image analysis algorithms sort of run to run or batch to batch, because there are batch differences.

Beth Cimini: And that’s, again, where that idea of like, well, is it am I now biasing the results every time that I change it, but I need to change it or it might be wrong next time, you know, you’re in this sort of like catch 22 of like, I want to keep it the same because I want my results to be comparable across batches. But also batches are different. So how can I sort of change it the exact right amount?

Grant Belgard: How do you approach rare phenotypes, subtle phenotypes, or like single cell heterogeneity?

Beth Cimini: Yeah, those are really hard. And those are things that largely as a field, we still haven’t, you know, come up with really good ways to deal with. I mean, one thing is always just kind of like, look at your data, look at your results, what level of results you can look at, you know, if you have millions upon millions of data points, you can’t look at them all one at a time, you can’t look at every image in something that has, you know, a million images, at least not without being there for three weeks. But figuring out how to sort of make sure things are, how you’re detecting the sort of rare things and how you’re making sure that you can even find them, that you’re not losing them in the segmentation process. Like I mentioned, you know, usually image analysis starts with, let me find objects that I care about, and your object detector might fail when weird biology is happening.

Beth Cimini: Hopefully, you know, something that sort of increases the likelihood of the weird biology, and you can go to look for it. But absolutely, I think we’re still only scratching the surface in terms of heterogeneity, in terms of how we deal with it, in the way that folks in like, the single cell sequencing field do in sort of large scale omics. At a smaller scale, certainly doing things like using super plots, where you can see all of the plots of your data, as opposed to like, not just like making a bar graph, bar graphs are the enemy. And sort of summarizations are the enemy can help with things like that. But it’s, it’s still something where I think we as a field need better methods. And I’m sure that, you know, smart people are working hard, and we will get better. But there’s still a lot of work to do there.

Grant Belgard: How can a team tell whether a model has learned biology rather than artifacts?

Beth Cimini: That is a great question. I wish I could answer for you. I mean, if there’s a particular piece of biology you want to, that you can induce that you know, you want to measure, then things are sort of like relatively straightforward, right? Because you sort of say, all right, I know that I should be able to induce, you know, a change in this marker, I have an assay that sort of like, you know, I have an antibody for that marker. And when I sort of add the drug or use the mutant that sort of where that marker should go up, it does. And again, doing things where you’re making sure that you’re careful to the person who designs the analysis is maybe blinded, you know, you’re not making sure that you grew all of your controls on, you know, last week, and you do all of your treatments this week, you know, basic experimental design stuff.

Beth Cimini: But especially when it comes to deep learning models, and especially when it comes to deep learning models, where we don’t have tons of data, you know, deep learning is going to learn exactly what it wants to learn. And your ability to tell what it learned can be really hard to figure out. There was a really great paper a couple years ago from somebody who had done a lot of work on cell painting stuff, and then use a tool called GradCam, they were sort of trying to classify cells with deep learning, and they use a tool called GradCam, which when you’re classifying a picture, it will sort of highlight in bright colors, the parts of the picture that it’s using to classify whatever you’re classifying. And it was saying the background is super important, not actually the cell.

Beth Cimini: And so in that case, they were able to check, they were able to sort of like, do this colored annotation of like important parts of the image. But when we teach, we have a bioimage analysis boot camp, we run a couple times a year. And we spend a couple of hours talking about how machine learning, you know, in general, and deep learning in particular, will lie to you, will learn what they want to learn, not what you want them to learn. And how can we out lazy the assay to try to figure out exactly what the model has learned? One of my favorite sort of like points to do is say, you know, what is the easiest way to train a classifier that you show it a picture of a cell and you ask it if the cell is interphase or mitotic? Just say interphase, because about 95% of cells, it turns out, are an interphase, and you’ll be 95% right. And most of us would take 95%, like 95% is a good day.

Beth Cimini: So if you just say interphase, and you don’t even interact with the thing at all, you’ll be right 95% of the time. And so trying to come up with counterfactuals, trying to sort of figure out what are ways the model could cheat, and design that into whatever algorithm you’re building, especially anything based off of machine learning are really critical, because it will cheat if you allow it to if it will find a way to do it. And we’re starting to see this now, I know, in some papers where people have taken LLM models, where it’s like, you give it an image, and it sort of tells you, like, this is a radiology image of such and such, and we see this condition in it. It turns out, if you don’t give them the image, they still say the same things. Like you you trick it into thinking it has the image, but it doesn’t, it will say the same exact things, whether it has the picture or not.

Beth Cimini: And so we need to be really careful to sort of make sure that we’re learning what we think we are.

Grant Belgard: Switching gears a bit, how do you balance tool building, tool maintenance, training, collaboration, and research? Beth Cimini Oh, gosh, I’ll tell you when I figure it out. I will say I have two fantastic team leads in my lab. Nodar Gogoberidze is our head of software engineering, and Erin Weisbart is our head of image analysis and training. And without them, you know, my life would be so much harder, and they sort of manage their respective domains really well. But yeah, there’s, and we have super smart, super hardworking people on the team, but it’s always a balance. You know, you want to be coming out with new features, but you also want to make sure that things don’t break. One of our software projects, BiLayers, is about sort of trying to promote containerization in software, because software containers are way less likely to break and are a way more portable artifact.

Grant Belgard: And so people might have to spend less time, you know, doing building, doing maintenance, because, you know, the things that they’ve made can be maintained more easily and for longer. It’s really a challenge to know, like, where in all of those things, like, am I spending my day doing the most good and helping the most people? So, so far, it’s just try a little bit of everything. But if you don’t build it, they won’t come. But if you do build it, and they don’t come, why did you build it? So you kind of need to be pushing on all of the pieces of the thing. And that’s why I like to describe it as a table. If we don’t have all of the legs of our table sort of in balance, the table isn’t a good table anymore. What makes scientific software sustainable? Oh, gosh, it largely isn’t. I will say scientific software sustainability is really hard.

Grant Belgard: And that’s a lot of it is because the vast majority of funding mechanisms are not aligned towards keeping things working. And I understand it. Like, if you’re like, well, would you like a shiny new thing? Or would you like to keep the thing that you have already working? You know, people want a shiny new thing. I totally get it. But at the same time, you know, you talk about a software project like the ImageJ project, which friends of ours make, which has about a million users a year and runs off of like a couple of people. And, you know, they’re always looking for ways to sort of fund that team better and be able to do more with that team. And this is something that a million people a year use. But nobody wants to fund something old when they could fund something new.

Grant Belgard: So I think that’s something that we as a sort of scientific community need to work on treating software as infrastructure. And we could fund infrastructure better in this country also and in the world also. But treating software as infrastructure and sort of saying in the same way that we know we need to maintain our roads and our pipes and our things in the digital age, software’s infrastructure too. And there should be more set asides to sort of keep stuff that way. You know, I mentioned containers earlier. That’s another way to at least make sure that things sort of can be used for longer. But new operating systems come out, new versions of languages come out.

Grant Belgard: And if existing tools don’t have funds, which funds equal time, then you end up with things like the heart bleed bug where, you know, open the SSL was being like done by one guy as a volunteer and something that half the internet relied on, you know, was insecure and nobody knew. So I think that’s something that we as a society could do better. And a couple of funding agencies are now starting to work in this space. But it would be great to have, you know, more focus on that being critical.

Grant Belgard: How can institutions build better career paths for bioimage analysts?

Beth Cimini: Yeah, I think first of all, they need to build career paths, not just better ones. Although I will say this seems like an area where the tide is turning. One of the first sort of big groups for bioimage analysis was the Network of European Union Bioimage Analysts or NEUBIAS, which was started in, I believe, 2015. And that was sort of one of the first times that people were like, hey, there is this career bioimage analysis. It exists. There are a few of us, like, let’s get together. And a few turned into a few hundred. There’s now a global bioimage analyst society, GloBIAS, that is trying to connect people from all over the world. But it’s a really strange career path in that, you know, it’s sort of, it’s by its definition, multidisciplinary.

Beth Cimini: You have people who started in biology or computer science or physics or math or, you know, something like that, and then pick up some of those other pieces to come to this, like, interdisciplinary space. And so it’s hard to sort of make a career path for it. But saying, hey, we have tons of our biologists making biology data, you should make sure it’s really quantitative and useful, is a thing that it now seems like more and more core facilities that I’m aware of are starting to have bioimage analysts on staff, which is, you know, a great step one. We ourselves have been running for a few years now a training program in bioimage analysis for postdocs, where we take awesome folks from the wet lab and train them in bioimage analysis. But I think definitely the sort of demand for help with quantifying images currently outstrips the supply.

Beth Cimini: But I think, you know, encouraging multidisciplinary rules, I think encouraging quantitative thinking in biology degrees, which having looked at lots of biology curricula from all over the country, you know, many biology PhDs still don’t require any statistics or computer science. And I’m not saying you need to be able to be a sort of like elite programmer when you leave, but being able to sort of have quantitative thinking be a really critical part of biology. I think there are a lot of fantastic biologists who got into biologists because it was the least math heavy science. And there’s fantastic biologists, and they do fantastic biology.

Beth Cimini: But I think as we go more and more into quantitative biology, saying, even if you’re not going to ever write a line of code, how can you structure your experiment to make it the most quantitatively interpretable is something a lot of people are never formally trained in, and they sort of pick up along the way. So I would love to see more biology programs sort of doing things like that.

Grant Belgard: What kinds of questions do people ask repeatedly? And what do those repetitions teach you?

Beth Cimini: Yeah, I think one of the most common questions that we get, so there’s an online image analysis help forum called image.sc, which if you ever, like, if you know nothing else about bioimage analysis, if you remember nothing else that I say today, remember image.sc. It is the central, like, help forum where you have more than 60 open source tools have one central help forum. And it’s a really friendly place to go and say, like, I have a picture, I want to know this about it, like, please, can somebody help me? And probably 10 people will. But we spend a lot of time, you know, sort of when we’re writing grants and things like that, chasing the hard problems, like the things that are currently really difficult to solve, you know, how can we do huge light sheet microscopy data that’s terabytes in size?

Beth Cimini: And one of the most common questions we get is just like, how with sort of really small data, can I determine what fraction of cells are positive for a particular marker, which is one of the easiest sort of image analyses to do, but people just sort of don’t even know where to start. And so I think we, it teaches me anyway, that like, we need to push the envelope on like the methods to solve things that are currently unsolvable, but that this education component of like, making sure that people know how to do the stuff that is computationally solved, but not necessarily solved for the person who needs to solve it and making it so that the tools are easy enough to use and the education is out there that people know the tools exist and know how to use them. We haven’t finished that part of the work.

Beth Cimini: And we can’t only focus on making the hot new tools because we need to make sure that the stuff that people are doing in their day-to-day lives, which is not necessarily the super hard stuff, is doable.

Grant Belgard: To talk about you, how did your own scientific interests change over time?

Beth Cimini: Yeah. For me, it really sort of began and ended with microscopy. When I was an undergraduate, sort of looking for a research lab, I was like, pretty sure I liked biology and wanted to do some research. I talked to a couple PIs and my undergrad PI, Bill Eldred, when I was interviewing with him, like pulled out a picture that they had taken of, I think this was a turtle retina, but like, you can just imagine this sort of like, brightly colored, lots of little specks all over the place, gorgeous image. And I was like, oh, that, that’s what I want to do. Like, I want to make those. And it started just as like, I wanted to do the microscopy and I still, I miss doing microscopy more than I miss anything else from the wet lab. But I got to graduate school and I wanted to answer a really fiddly, you know, thing. I studied telomeres, which are the caps on the ends of your chromosomes.

Beth Cimini: You have 46 chromosomes in each of your cells. And so you have 92 telomeres in each of your cells. And I wanted to say, you know, depending on how long the telomere is, so I have to measure that with one color of microscopy. There’s a protein that comes in two splice forms, two flavors, you know, does the length of the telomere change how much of protein flavor A versus B is there? So I have to measure three different colors very quantitatively in very small spots, a hundred of them per cell, hundreds of cells many times. And counting it by hand all of a sudden was not going to work anymore. So I had to learn to code and I really didn’t want to.

Beth Cimini: But once I did, I found the sort of like puzzle solving aspect of that to be so satisfying that I love doing bioimage analysis, I sort of picked up to the despair of my PhD advisor, all sorts of little side projects, like helping my friends analyze their data. And that when I found out when I was close to graduating my PhD, like this is a job, like I’m like, this is a job, like this thing that I love to do for fun. So I realized that collaborating with other people and helping sort of solve these image analysis puzzles, like was a huge joy to me. And that brought me to the Broad. And now I never dreamed of writing software, though, like our software that other people would use, I’d written some own code, you know, for some own software myself, I realized how much I personally enjoy and how much I personally get out of like making stuff that makes other people’s lives easier.

Beth Cimini: Like when I was a grad student, nobody was going to care at the end of the day, like what my project turned out, probably. But if I make tools that like make other people’s lives easier than every day, I’m excited to come to work, because I know that the work that I’m doing is going to lead to somebody else being able to make a cool discovery they wouldn’t have been able to before. And that for me is a way better reason to get out of bed. And a way better reason to come to work and work really hard.

Grant Belgard: What’s changed as your work has shifted from individual projects to leading people in programs?

Beth Cimini: I mean, it’s definitely now more of a, you know, you have to be thinking not just about how am I going to get through the next two or three weeks, but I have to think how am I going to get through the next two to three years and sort of really be planning for the long term because grant cycles are, you know, about a year long, you have to apply about a year before any money comes in, and you have a team of people and you want to make sure that you can keep the great people and pay them. You know, we can’t pay, especially our sort of computational folks, what they truly deserve, but we pay them as best we can with NIH budgets being what they are. And so really just sort of having this sort of like multi-level contingency plan of like, well, if I get this grant, then we’ll do this. But if we get this other grant, we’ll do that. And yeah, it becomes much more than just about you.

Beth Cimini: It’s about your whole team and how can you make sure that every single one of them has what they need. And so you can’t just be thinking about what we need now. We need to be thinking about what we need to do now to make sure we’re good a year from now, which is higher stress than just sort of thinking about what I need to do in the next three weeks. But the folks I get to work with are the best in the world. And I love getting to feel like I’m making their lives easier.

Grant Belgard: In your current role, which skills have you found matter more than you expected?

Beth Cimini: Multitasking for sure. And like fast task switching. The other thing, and it’s not a thing that I’m naturally very good at, is just documentation. Like leaving things in a place where if you need to task switch, you need to come back to something two or three months from now, like you actually remember where you were and what you did. And like, I was never the person who like loved writing in their lab notebook and sort of carefully documenting everything. But it became, you know, do this or fail, which was sort of how I came to coding too. So I guess I’ve had a lot of skills that it’s been like, well, you’re going to get better at this or you’re going to be bad at your job. So documentation was one that I really had to get much better at.

Beth Cimini: And I’m lucky to work with, I mentioned Erin Weisbart at my team already, but also Anne Carpenter are two of the best, like organized people who make the best documentation that I know. And they’ve taught me a lot as I’ve worked in this job.

Grant Belgard: Switching now to, to advice. What advice do you wish you had heard earlier? Or at least heeded earlier?

Beth Cimini: Yeah. I am grateful for all of the sort of like side projects that I got to do during my PhD. I wish I had realized earlier that the fact that I only enjoyed my side projects and not my main project meant that like a biology research career where I study a particular biological problem was not probably where my brain was happiest. And really not feeling like because that is the normal path, that is the path you must take. You know, I had very definitively decided I wasn’t going to be a PI and I came to the Broad to be a staff scientist. And then one thing led to another and ended up being a PI of a very different kind of lab, a lab where we do computational stuff. But I was, I had decided that these were the normal paths and therefore they’re the only paths available to me. And now I’ve been on this very unusual, strange own path.

Beth Cimini: And I’m happier than I think with any of the sort of normal paths, but it can be hard to imagine something besides what you’ve seen. So if you hate the things that you’ve seen, like go look for more stuff. There are weirder ways to get to a place where you’re happy than, than you possibly ever knew. You just have to find the people who’ve been on those strange paths.

Grant Belgard: I think I know the answer to this, but, you know, we always like to echo such things, right? So at what point in an imaging project should someone seek image analyst input?

Beth Cimini: Yes, early, as early as possible, preferably before you’ve done, once you’ve done your first pilot, and you should definitely be doing pilots, pilots with controls. We have a graphic in one of our papers. It’s a PLOS biology paper from 2023. The first authors are Senft and Diaz-Rohrer, where we show the circle of bioimaging and bioimage analysis. And they should be a circle. It should be that one thing feeds into the next, feeds into the next. And that your data, your image analysis and your data analysis helps you design the next best experiment. And that when you do your pilots, you analyze them all the way through to make sure that you can actually detect the statistical measurement you want to make. Otherwise, you don’t know if you can actually detect the thing you care about or not. Teams like mine have office hours.

Beth Cimini: A lot of places, your friendly imaging core will have a friendly local image analyst. If not, GloBIAS, the Global Bioimage Analyst Society, has a database of people who have agreed to be contacted on their website that you can just reach out and find a bioimage analyst near you who works on things that you work on. Because the sooner that you figure that out, the more likely that you’ll never end up sitting in a bioimage analyst’s office and then saying, the information you want just isn’t there. Those are the days I leave work the sad. It’s just when we have to tell somebody, I’m sorry, the information you want, the way the experiment was designed, we just can’t say anything.

Grant Belgard: So many parallels here with sequencing analysis.

Beth Cimini: Yeah. It’s so tricky. I hate having to do that, but sometimes the data just isn’t there. So the sooner you talk to us, the less likely that you’ll ever have that conversation.

Grant Belgard: What mistakes should people try hardest to avoid when collecting image data?

Beth Cimini: As we get more and more into people using automated microscopes, I think there will be less of this, but going in with a preconceived notion, you have to know a little bit. You have to know what stains and stuff should be present. But if you go in and you decide that you’re only going to take pictures in your controls of cells that look like this, and then you’re only going to take pictures in other fields that look like something else, you’ll find a difference. But it might not be the real underlying biology. So try to design things so that you can find more than just the preconceived idea you came in with. And it can be hard to do that while also making sure that you can find the thing that you know you care about, but it will help you sort of avoid some of those biases that will doom you before you can get started.

Beth Cimini: And that’s where I mentioned something like having your friend plate your samples for you so that you don’t actually know what’s in each well until after the experiment’s over.

Grant Belgard: How can early career scientists make invisible infrastructure work visible?

Beth Cimini: Oh, gosh, I think this is a thing that we as a sort of field still need to work on. But I mean, I think there are now more options for publishing things like the Journal of Open Source Software, like academic currency still runs on citations. And so making things citable, you know, make things like Zenodo DOIs, at least like put digital object identifiers on your work. So you can say, look at these things I made, they all have a digital object identifier. You know, some of it is, you know, administration has to meet us halfway. And you know, people who are hiring have to sort of, say, it’s not just about the papers, it’s about maybe what the papers do. But, you know, playing the game to sort of like get the metrics that you need.

Beth Cimini: Well, also, we as a community work to make it so that things like, you know, GitHub stars, or, you know, usages of tools like page visits to a blog that you wrote, that actually really helps people figure out how to do something like none of those are traditional academic currency, but all of them might be really valuable.

Grant Belgard: How should someone decide whether bioimage analysis could be a good career fit?

Beth Cimini: I think it’s really well suited to puzzle solvers. So if you love solving puzzles, I think bioimage analysis is a great career fit. I think it’s a relatively easy thing to just do, because there’s tons of images online in places like the Bioimage Archive or the Broad Bioimage Benchmark Collection. There’s tons of free tools, like, you can just sort of play and see if it feels like this is a fun thing for you. If you’re going to do it as a career, you probably have to be somebody who loves working with other people, if that’s the thing that excites you and not drains you. But I think if you like solving puzzles, and you like working with others, those to me are, and you’re organized and can sort of track having multiple things going on at once. Those for us are the things we hire for in bioimage analysts for helping them succeed.

Grant Belgard: Which bottleneck, if solved, would change bioimage analysis the fastest?

Beth Cimini: That’s a good question. It’s going to be a very, like, unexciting answer. But like, metadata standardization. I mentioned there’s like a million different things that could be in an image, and a cell could be one pixel, it could be the whole field of view. We have no common way to sort of describe what’s in an image, which means we have no common way to even say, like, should these two images look the same? Yes or no. And that goes to the thing about quality standards being hard. If we just had a common descriptive language, and smart people are working hard here. But if we all just agreed on how to describe the things, then, you know, having big databases of images and being able to make cool deep learning models and sort of suggest how your analysis should go would be so much easier.

Beth Cimini: But right now, you have to tell me and show me and we have to have a consultation and sit down and I love doing those. So I kind of don’t want to automate them away. But we could automate them away and make this a lot more approachable for everybody if we just agreed on a common way to describe things. It’s a very uncool answer, but it’s really what we need.

Grant Belgard: What worries you the most and what excites you the most about the next few years of AI in this field?

Beth Cimini: Yeah, I think I’m going to distinguish a little bit here between like deep learning in general and like generative AI, you know, large language models particularly. I think deep learning for bioimage analysis has proven incredibly powerful, especially in the context of segmentation. And people are doing fantastic work there. You know, we’ve got some papers out of like deep learning models that help you get more information about images. And all of that is great. Where I worry the most is large language models, at least the ones the common commercial ones that people use have been trained are designed to give you confident answers always. And bioimage analysis is a space that is just full of nuance. Very small changes in experimental design or in what you want to find make huge differences in workflows, make huge differences in the sort of measurements you might want to make.

Beth Cimini: And if you look at the where existing benchmarks are for bioimage analysis, and there aren’t many, LLMs are terrible at them. But because they’re advancing in other fields, because people like them in other fields, I worry about people confidently adopting tools that are not right for the job, without even realizing, doing their best, not saying like, well, this is wrong, but I’m going to do it anyway. But just saying, oh, well, the LLM says it’s confident that this is right. And I don’t think that the tools are there yet. And I think we need to understand it’s a very complex problem and not lose our skepticism. Because, you know, next probable word token predictor, you know, tells us that this is right. And those tools are powerful, people are using them to do cool things. But we need to recognize their limitations and not just sort of trust what they say.

Grant Belgard: What would you like the field to look like 10 years from now?

Beth Cimini: I want there to be slightly more professional bioimage analysts. I still think we need more of them because I think there’s still a lot of hard unsolved problems. But I do hope that a lot of the people who are currently stuck on the easy problems, which I mentioned, are actually like the majority of the problems, find it so that they can understand bioimage analysis themselves, they can learn it themselves, they can do it themselves, and then they can move on to something else. And maybe some of them will become the professional bioimage analysts who are working on the hard problems. And I think as microscopes continue to get better and better, we will come up with new hard bioimage analysis problems. But I hope that for the biologist who just wants to analyze the pictures and get on with their day, they can do that a lot more easily.

Beth Cimini: I think there’s, you know, again, we’re working on the table to make the tools better, to make the education better. And I hope over the next 10 years, I can look back and say, wow, things are a lot better than they were in 2026.

Grant Belgard: Finally, what should people remember from this conversation?

Beth Cimini: I hope what they remember is that microscopy and image analysis, like, are really powerful because they’re diverse. But therefore, it’s a discipline that’s full of a lot of like subtle pitfalls. And it’s good to talk to experts about it. But that the experts are out there, we exist, and we’re kind nerds who love talking about this stuff. So find us on image.sc, or find us in the GloBIAS Database, or come to my team’s office hours. We’re out there, we want to help you. And we really love talking about this stuff.

Grant Belgard: Thank you so much for joining us. It was a lovely conversation.

Beth Cimini: Yeah, thank you so much for having me. Have a wonderful rest of your day.

Transcript of Episode 88: Ben Langmead

Disclaimer: Transcripts are automated and may contain errors.

Grant Belgard: Welcome to the Bioinformatics CRO Podcast. Today, we’re joined by Dr. Ben Langmead. Ben is a professor of computer science at Johns Hopkins University, where he directs the Langmead Lab with a joint appointment in biostatistics at the Bloomberg School of Public Health. His group works at the intersection of computational genomics, sequence alignment, text indexing, statistics, and high-performance computing, with a focus on open source tools and resources that help life scientists use high throughput sequencing data. Many listeners will know tools and resources associated with Ben and his group, including Bowtie, Bowtie 2, recount3, and Snaptron. He is also founder and principal of InOrder Labs and has made a large collection of teaching materials freely available online. Ben, welcome to the podcast

Ben Langmead: Thank you for having me. It’s an honor, Grant

Grant Belgard: So for listeners who know your name but not your day-to-day, what are the biggest problems taking up most of your attention right now?

Ben Langmead: We always like to work on problems where there is a hole in genomics everyday practice that can be filled with something from computer science. So in the last, six, seven years, I would say that the main thing that we’ve been working on is pangenomics. So we’re in this new era where long read sequencing is much less expensive than it used to be, the reads are higher quality than they used to be.

Ben Langmead: And so we can basically make our dreams come true in terms of assembling many new genomes. And we have our new telomere to telomere human assembly, and we have the Human Pangenome Consortium producing high quality human genome assemblies in bundles of hundreds at this point. And so this is just one of the many ways in which in genomics, we would love to move in this direction of more should be better, right?

Ben Langmead: More is better, right? But the truth is usually the more comes first and then the better maybe comes later if we really work at it. And when it comes to pangenomics, I think we’re still working at trying to make a typical everyday bioinformatics workflow truly benefit from the existence of all these very high quality reference genomes.

Ben Langmead: And to make that happen, there’s been a really superb, fun interplay between computer science and genomics, right? We needed more when it came to we need better indexing approaches, we need better algorithms, and we need it to not cost much more than it used to cost us to analyze data with respect to a single reference genome.

Ben Langmead: And computer science has been providing those missing pieces, and in fact, we now have multiple choices for where in the computer science theory literature we can turn to find new solutions to these problems. And so we’ve had a lot of fun since roughly 2019 going down multiple roads, but mainly down this road of compressed full text indexing.

Ben Langmead: How do we fit pangenomes in efficient compressed full text indexes? And then how do we let that be the basis for the next generation of tools, whatever’s coming after the current generation, but that can allow people to work seamlessly with huge pangenome collections.

Grant Belgard: How do you explain your work differently to a computer scientist, a biologist, and a clinician?

Ben Langmead: Yeah. So it’s funny because a lot of the work I do, it’s definitely more applied than what a typical theoretical computer scientist would work on. And in fact, theoretical computer scientists look at me and they see someone that’s almost like a biologist.

Ben Langmead: And which a biologist would chuckle at, right? I’m not even remotely a biologist. But to people doing very theoretical computer science, the fact that I understand the domain and understand the application of these things makes me a little bit more foreign.

Ben Langmead: And so I do, when I’m speaking to computer scientists, I tend to speak in terms of these are the open problems, and I try to frame it like it is a computational problem. The open problem is we need to be able to build a full text index, but it has to be not much bigger than the full text index for a single genome, blah, blah, blah.

Ben Langmead: And then for for people all the way on the other side in biomedicine, what I really want them to appreciate about the work we’re doing is that it should be freeing for them. It should make their lives easier. It should reduce friction. It should address the things that are bothering them about the computational work they’ve done in the past, where they say, “Oh my gosh, we’re doing all of this with respect to a single linear reference.

Ben Langmead: I know that I missed the– I know that there are entire genes that are not present in that reference. I know that there are rare alleles that are present in that reference, and yet I’m making all my arguments based on that reference. What, this is bothering me.” And so what I want the people in biomedicine to understand is that all the ways in which you feel like there’s a closed door it could be opened if only, we had an interplay and talked about what really would you prefer to do?

Ben Langmead: You mentioned at the top, our work on I’ve been talking about pangenomics, but you mentioned recount3 and Snaptron. That’s one of my favorite stories because recount3 and Snaptron, those are large scale summaries and indexes built over public RNA sequencing data sets. So these data sets, the raw data is already out there in the public.

Ben Langmead: It’s in the sequence read archive and in other resources like that. And we made summaries of that data available because it’s much easier to use than the raw data. And then we built indexes over it in the hope that we’re producing something like Google, but for archived RNA sequencing data.

Ben Langmead: And when we started publishing papers I heard from one of my colleagues at Hopkins Jonathan Ling, who wanted a meeting, and when we met, he said, “We love your paper, but it’s not what we wanted. What we wanted was this.” And so Jonathan did me the favor of making the link. He saw what we were doing, but he also saw it wasn’t what he wanted.

Ben Langmead: So he got in touch and told us what would be the missing pieces that would take our work all the way to where he was working. And so I think when the system works well there are people in each of these silos, right? The super theoretical computer science, the in-between computational genomicists like me who are maybe like glorified brokers, and then people on the biomedical research side, all understanding just enough of the other people’s language that they can phrase their problem.

Ben Langmead: And so though when I interact with people in biomedicine, I feel like my job is to try to make them understand which are the doors they thought were closed or like which are the problems they thought were too hard that if only we could solve a problem in computational genomics, they would suddenly be open and that can be more creative, right?

Ben Langmead: You can go back, ” tell me more about how you even framed your problem in the first place. Did you do, did you look at the public data right away or did you wait to look at the public data until the end to validate what you were trying to do?” So absolutely we all have different languages and we all talk to each other.

Ben Langmead: We all talk to each other in our own language, but we do especially well when we learn a little bit of each other’s language. And right now I would say those three pieces are talking to each other pretty well and I think better than when I first started in the field, which was when second generation sequencing was still young.

Grant Belgard: What has changed most about the kinds of questions that people ask of sequencing data since you started in the field?

Ben Langmead: Yeah. When I first started, second gen was really– we were really just starting to see data. So when I started grad school in 2007, and I think some of the people I was working with at University of Maryland, they had just seen for the first time a FASTQ file from Solexa or a FASTA file or something like that full of reads.

Ben Langmead: And it was just starting to sink in that they truly could not make progress with this new data type unless there were new advances in computer science. There was no hope of putting these things in BLAST. It just wasn’t gonna work. And so the problem at that point was it was almost existential.

Ben Langmead: It was how do we even make progress with this data type? But there was very fast progress. Computer scientists and others did snap to attention, and suddenly we had MAQ and we had Eland, and then there was BWA and Bowtie, and suddenly we were off to the races and the problems moved downstream, right?

Ben Langmead: So the next question was, “Oh maybe we can use sequencers to study other things. We can sequence RNAs.” And so the “what do we even do with sequencing data” sequence of questions have been pretty well answered, and we do have pretty good algorithms. We even have, what you might call like best practices and best practice workflows, and we have benchmarks, and we have comparison papers and bake-offs.

Ben Langmead: And so there’s been a lot of, I think, flourishing of improved tooling and an improved understanding of what it is you’re supposed to do when you first start with a sequencing data set. But now I’d say we’re changing our attitudes in a couple of different ways. So first of all, I think that we could change our mind about what we even use sequencing to do versus what we use sequencing data that’s already out there to do, right?

Ben Langmead: In the same way that if we start a new project today, or if not even a science project, if we’re doing a renovation to our house or we’re doing s- we’re doing something difficult today, we probably need to start to investigate things like who are the best vendors and what are the tax implications, right?

Ben Langmead: And we Google a thousand things or we ask ChatGPT a thousand things. Similarly in science, there is all this archived data, we could start there. It’s not typical, but we could, right? So in other words, even upstream of deciding what it is you want to sequence, you could decide I want to narrow my– I want to form my hypothesis or I want to narrow my hypothesis by looking out in the public data world, and we have, many petabases of of public data in the sequence, Sequence Read Archive, for example, and letting that be a a wellspring for new hypotheses or narrowing our hypotheses.

Ben Langmead: That is, I think, still a little bit of an undiscovered country, right? We, w- it’s still the case that the Sequence Read Archive is a little bit un– not, I won’t say unused, but it’s a quiet place, right? It’s not some-something that people are querying every day and learning from every day, even though we’re paying good money to have all that data out there.

Ben Langmead: Then the other way in which I think attention is shifting is that in order to build those pipelines, we made some bargains. We decided we were okay with certain trades, and I think we’re revisiting some of those now. And so all the work in pangenomics, in a sense, is really about addressing one of those maybe not so good bargains we made at the outset, which was to use a single linear reference, and that gets us very far.

Ben Langmead: The single linear reference is really, has been good to us in many ways, right? It defines a very easy to understand coordinate system. Anything you might wanna refer to is a nice, simple 1D interval. You and your friend can refer to the same interval of the genome in exactly the same language and no one’s confused.

Ben Langmead: But on the other hand, it brought us reference bias, which we’ve been living with for, a decade and a half now, where when we use a single linear reference, there are places on that single linear reference that are not a good reference for the individual we’re studying. They’re just not similar enough to the individual we’re studying, and it creates confusion from then on, and there’s often not really a way to recover from that confusion later.

Ben Langmead: There can be in some cases, but… And so I think that’s another thing that we’re turning our attention to. We built wonderful tooling, but it was built on some trades, and now that we wanna address those trades, it does require a bit of uprooting of how we did things before and a bit of disturbance so that we can adapt.

Grant Belgard: How would you explain sequence alignment to someone who uses aligners but has never thought about how they work?

Ben Langmead: Yeah that’s funny. I remember after Bowtie was out, I went to a conference, and at the conference was my colleague from Penn State, Anton Nekrutenko, who was one of the founders of the Galaxy Project. And he was saying, he basically asked me that same question because he said, “People are asking me what is the tool that is the fastq to SAM converter?

Ben Langmead: And I keep telling them it’s Bowtie, but it’s not really a converter, right? It actually has an algorithm in it.” Yeah. So I do explain this in the classroom once a year every year. And we start with the basics of, why it is that the sequencer emits fragments.

Ben Langmead: Why wouldn’t the sequencer just give us what we want, which is the complete picture from beginning to end with no mistakes? And, my analogy is it can’t do that. It’d be like you can read, but if I gave you a book and told you to read the whole book aloud from beginning to end without ever stopping, you wouldn’t be able to do it, right?

Ben Langmead: You would have a coughing fit, or you would have to go to the bathroom or something like that, right? So similarly, sequencers can’t do that, which means we’re inevitably faced with this problem of having fragments that have to be pieced together into the completed picture. And we don’t have to do this with alignment.

Ben Langmead: We can also do it with assembly, and so there’s… So I need to explain the difference between those two things, and then explaining alignment is it’s a bit like putting together a jigsaw puzzle, but where you can see a picture of the completed puzzle. And so you can hold the pieces up to the puzzle, and you can figure out how to piece it together that way.

Ben Langmead: And all of that is all well and good, but it doesn’t tell you what any of the algorithms are. So then we move into how do you do exact matching? And that’s, many decades of computer science worth of stuff we can talk about, and including some fantastic algorithms that we don’t generally use in genomics but that are used elsewhere on your computer in your day-to-day computer use.

Ben Langmead: And then we say, “Ah, but there can be differences because of sequencing errors and genetic differences.” And so then we go into approximate matching and then all our, beloved bio sequence analysis tools like Smith-Waterman and Needleman-Wunsch and such. So I guess I usually explain it through the storytelling of what are the problems we need to solve, and then why do we need this new kind of algorithm to solve these different problems?

Grant Belgard: Why does indexing matter so much in genomics?

Ben Langmead: Yeah. So indexing was in some sense the answer to the question earlier of how do we even… So like when the Solexa data was coming over the wall and people could see how voluminous it was a lot of the existing methods BLAST included, were all about scanning the reference database looking for good matches for the query, right?

Ben Langmead: You paste your query sequence into the box, and BLAST will then go and it does this in a fast and distributed way, but it’s actually literally scanning reference sequences in order to answer your question. And I think once Solexa came along, we outgrew that strategy because we had so many queries, it didn’t make sense to, for each of them, go scan the reference.

Ben Langmead: The analogy I like to use is your search engine, Google. When you put a query into Google, it is not literally scanning all those web pages while you wait and then giving you an answer because that would not be as instantaneous as you need it to be. It’s using an index, right? So indexing comes into play when you have one reference that you know you’re gonna use again and again.

Ben Langmead: I’m gonna go searching for a needle in this particular haystack over and over again. And not only that, but I’m gonna search for many needles in that haystack. When you get to that point, then it becomes a, a question of amortizing. In computer science we like to amortize effort.

Ben Langmead: So if we’re gonna put in effort ahead of time to build a data structure to help us solve this problem, it makes much more sense to build it over the haystack. And so building that data structure to help us search inside a very large collection of text that’s what we call text indexing.

Ben Langmead: And the name index is a good one because it’s a good analogy. The index of a book is a perfectly good analogy for what it is we’re doing, right? We’re building something that has pointers into the huge the vast amorphous data set that we’re trying to index. It has pointers into it that help us jump to exactly where we need to go at query time.

Ben Langmead: And so text indexing has, in computer science, has been around for a very long time. There’s a wonderful sort of parade of data structures that came out over time, and genomics has absolutely been on the bandwagon of following and adopting each of these as they became popular. So suffix trees and suffix arrays were already very popular when I started graduate school, and then through the work of others and through my work in graduate school, these more, I won’t say modern but very frequent, very frequently used today this very frequently used idea of the Burrows-Wheeler transform became the one that really was useful because it was the one that let everything fit in memory on a typical computer at least at the time.

Ben Langmead: And these days now the trick is trying to use some of those same tools to get entire pangenome indexes to fit in memory on a typical computer.

Grant Belgard: How has your thinking about alignment changed as sequencing technologies have changed?

Ben Langmead: Well, sequencing technology, the way it has changed it didn’t have to do this, but it did in fact reveal new problems to solve as it went. I should say yes and no, because sometimes it reveals new problems and sometimes old problems become current again. So for example, long-read sequencing produced reads that were more amenable to the kinds of alignment algorithms that had already been developed for Sanger sequencing.

Ben Langmead: And therefore part of the need from the computational genomics community was let’s rediscover and dust off some of those older algorithms that we had for longer reads. So sometimes it points us to something we’ve already done, but sometimes it shows us that something entirely new is required.

Ben Langmead: And so these days, in a way, the advent of long reads has changed everything we want to do, whether it’s with short reads or with long reads when it comes to alignment, because the long reads gave us this abundance of new references. And the trick is: how do we use them? How do we use them in a way that truly benefits us, but for not much more cost?

Ben Langmead: And so that is a truly new… It’s in one sense, a truly new way of viewing the problem that we have to solve. But the good news is it actually calls on some of the exact same methods that we were already using, right? So like the Burrows-Wheeler transform, which I mentioned before, this is like the core.

Ben Langmead: If you’ve used BWA or Bowtie, the Burrows-Wheeler transform is at the core of those tools. It’s at the core of some other genomics tools as well. It has many variants like the positional BWT and things like this. But the BWT was originally invented not for indexing, it was invented for compression. So it was invented to basically compress files on your computer, like Zip does.

Ben Langmead: Which is a fortunate thing because this problem of indexing pangenomes, it requires compression, right? One thing that won’t work is just building the same old indexes over each of the sequences in your pangenome. Let’s say your pangenome consists of hundreds of human haplotypes, because that’s just gonna grow proportionally to the length of all those haplotypes, and you will very quickly exhaust the memory of the computer you’re using.

Ben Langmead: So it turns out, happily, that the exact tool that we’ve been using since, 2009-ish for our, to solve alignment problems, the Burrows-Wheeler transform, was invented for something like this in the first place. Compressed representations. So the latest revolution is really about compressed indexing and being able to efficiently query compressed indexes, so where the index itself is not much bigger than a fully de-duplicated version of the text that it’s indexing

Grant Belgard: Where do pangenome approaches feel mature and where are they still experimental?

Ben Langmead: That’s a tough question to answer. So I think that there are some things that people are getting used to doing that involve using multiple references. But whether people are yet used to fully pangenome approaches, I don’t quite know, right? I think people still wonder this, right?

Ben Langmead: They still wonder “how do I do this in my day-to-day?” They know there’s such a thing as pangenomes and graph pangenomes, and there’s graph aligners, which are probably the most mature single thing we can point to. So things like VG and Giraffe, these are quite mature, so people can use these tools.

Ben Langmead: But what people do day-to-day still looks to me like thinking of genomes as individual things, but trying to use more of them. So instead of embracing the whole forest, it’s like just swinging from vine to vine or something like that. So let’s try to use this reference genome, and if it looks like the situation is breaking down in some situations, like it looks like some of my reads failed to align, but they did align to this other reference genome, that means that other reference genome is the one I need to study this particular thing.

Ben Langmead: Maybe therefore I’m gonna hop from this reference to that reference and, but, and do everything again, but with respect to that other reference. Which isn’t even a very bad way of doing things because you, it doesn’t take a large number of reference genomes before one of them is probably a pretty good reference for whatever it is you’re trying to study.

Ben Langmead: But it’s definitely not yet the case that we train people to use pangenomes instead of linear reference genomes. So that’s, I think, one thing that eventually needs to change, is it has to be part of our training. We have to learn what they are and what reference bias is, how to detect when we might be falling prey to reference bias, and then understanding how to use the pangenome to cure that.

Ben Langmead: But also there’s a lot of unanswered questions, right? So there’s actually just plain still more work to do in pangenomics. And, I think what a lot of biologists would say is, “It’s great that you have the pangenome for me to use, but how do I explain it? If I use it in my paper, how do I explain what I found?”

Ben Langmead: There’s a anxiety about the notion of pangenome coordinates. Like, how do I even describe where I am in a pangenome? And is it gonna make sense to somebody who reads my paper? Compounded with that is the fact that pangenome coordinates themselves are a little bit mythical, right?

Ben Langmead: It’s not necessarily true that a pangenome does or should have a global coordinate system because it’s not necessarily globally collinear. If you insist on understanding a pangenome in a global way, the global landmarks are gonna fall away more and more the more genomes you add, right?

Ben Langmead: Because some genome is gonna disagree about what’s next to what in the genome, and as soon as you add that one, you’ve knocked out part of the common coordinate system. So part of what we need to address is maybe we don’t need to insist on the pangenome coordinate system being truly global.

Ben Langmead: Maybe it’s something that we navigate in a way where it would be awesome if we could answer our questions in a fully zoomed out global way. But when that’s not possible, we should zoom in and say, “Okay I can provide a good coordinate system for what it is you’re studying, but only if I limit my attention from here to here and discard these sequences from the database.

Ben Langmead: Now I can give you a really clear, nice picture, maybe like a multiple alignment or something like that of that area.” But these tools are also missing, right? In fact, at this point, the pangenomes are getting to be so big it’s hard to even build a multiple alignment of them, right?

Ben Langmead: We already have pangenome data sets for which I doubt we’re ever gonna have a good, reliable multiple sequence alignment of those sequences, even though they are as related as humans, for example. So another obstacle is how do we both take advantage of the additional data that the pangenome’s giving us, but without going too far, right?

Ben Langmead: So more is better right up until the moment that it’s not, right? More is better right up until the pangenome stops being a good global coordinate system for what you’re trying to study. So we do need a way of saying, “More is better, but there are limits. And if you wanna, if you’re studying what you’re studying, we recommend this subset.

Ben Langmead: Or in this area, you should be using this subset. Or I see what you’re studying I’ve detected that these 80% of the references are perfectly good for what you’re trying to study, and these 20% are just gonna be confusing, so I’m gonna zoom in for you.” There’s still a lot of work, I think, to be done in that area.

Ben Langmead: So I don’t know if it’s even time yet to train people to learn about pangenomes. It’s certainly a good time to learn about reference bias and how you can avoid it by moving from linear reference to linear reference, and there are certainly some problems we can already solve with graph pangenomes and graph aligners.

Ben Langmead: But the future is still fairly uncharted, I would say, when it comes to how are people really going to switch to pangenome thinking, right? We’re still trying to wrap our heads around what pangenome thinking is, and therefore how we should, what tools we should build and how we should train people.

Grant Belgard: How do metagenomic data problems differ from human genome data problems computationally?

Ben Langmead: In some ways they’re quite similar. So for example, in metagenomics, maybe you’ve collected a sample from the human gut or from a soil sample or a surface that you’ve swabbed in a hospital or a waterway, and you are sequencing everything that is there. So you are getting sequencing reads derived from the DNA genomes of whatever was there.

Ben Langmead: And again your job is to put together a puzzle. So quite similar to if you had just sequenced one human. But the problem with this puzzle is you don’t necessarily have the picture of the completed puzzle for all these sequences, right? Because there’s a lot of organisms that you could collect in a metagenomic sample for which we just don’t have reference genomes and maybe don’t even have nearby.

Ben Langmead: Maybe we don’t even have cousins, we don’t have anything close enough in our reference databases. So that’s one immediate difference. Another difference is it’s quantitative in the sense that if a organism is there in a higher abundance, you’re gonna get more data from it that usually when we’re sequencing, say DNA from a human, we want to sequence equally from all parts of the genome and that’s quite possible.

Ben Langmead: But for a metagenomic sample, we usually have to deal with the fact that some organisms are much more abundant than others. That’s another additional problem But other than that, it actually looks similar, right? Because we usually analyze these metagenomic sequencing reads with respect to a reference database.

Ben Langmead: That reference database now is not just one species of reference genome, but it’s a whole collection of species of reference genomes. But again, one of our most valuable tools in this situation is indexing. So this again is a text indexing problem. It’s even, again, a compressed text indexing problem because if you think of what reference genomes we have for what clades in the tree of life, we have a huge abundance of reference genomes for some clades and then poorly sampled reference genomes for other clades.

Ben Langmead: But like things that are important to human health, things like COVID and SARS-CoV-2 and E. coli and things like this we have tons and tons of reference genomes in those clades. But then in other clades, maybe for organisms that are hard to culture or hard to isolate, we don’t have as many reference genomes.

Ben Langmead: So we have some parts of the tree of life where it is crucial that we do compressed indexing because we’re just not gonna be able to fit all those sequences in unless we do. And so indexing, compressed indexing, these are still the valid problems. We end up using, in practice, often an expedient here, which is to build a k-mer index.

Ben Langmead: So in other words, the sort of keys, if you think of the index of a book, right? The keys there are the key terms that you might wanna look up. In a k-mer index, the key terms are just substrings of length K, right? So those are the key terms that are held in our metagenomics index. So that’s, it has its pros and cons.

Ben Langmead: K- Once you do this, you have to pick the value of K, right? So you have to decide what is the true one and only substring length that’s gonna tell me what I need to know. And, almost everybody just picks a default, which is often 31. But it is far from obvious that’s what should be done.

Ben Langmead: It’s also far from obvious that’s what should be done everywhere in the tree of life, because again, some parts of the tree of life are very dense with sample genomes and some are very sparse. Some parts of the tree of life just have more divergence among the individuals in that clade, and others have less.

Ben Langmead: So we use these expedients, and that’s because it’s hard. It’s harder than, for example, the human pangenome, because the different clades don’t have sequence similarity across the clades, right? So the compressed indexing is only going to be beneficial within a clade. It’s not particularly beneficial across clades.

Ben Langmead: So we have to rely on these expedients. We can’t quite do anything quite as comprehensive as a full text index. At least not usually, though some tools do this, and I think the future probably is for tools that do this. So it is in some ways a harder computational problem, but it is a similar one

Grant Belgard: What makes a bioinformatics tool trustworthy?

Ben Langmead: So the research world is appropriately a world where we like to let 1,000 flowers bloom and a lot of software gets put out there. And this is good, right? Students need to be able to have an open playing field to go ahead and make their contribution.

Ben Langmead: But not everything is a hit, right? And it’s impossible to predict what’s gonna be a hit. It’s slightly possible to predict what’s gonna be a hit, but generally it’s not possible to predict what’s gonna be a hit. So the truth of the matter is everything starts on kind of equal footing.

Ben Langmead: Maybe this is gonna have traction, maybe it’s not. And if something has traction, then hopefully the student and the advisor are seeing that happen and responding. And often the way you respond is you say, “Oh, wow, we should be thinking about releasing new versions and talking to some of the users, or at least trying to figure out what the users are thinking.”

Ben Langmead: And sometimes things get in the way of that happening. Maybe the student graduated and the PI is very busy and teaching, and so nobody notices. This is okay. But sometimes everything comes together and we can see, oh, this is a hit. Let’s put out more indexes, for example. Maybe this is an indexing-based tool and some people wanna use human indexes, but other people wanna use mouse indexes, so let’s put out more mouse indexes.

Ben Langmead: Or maybe people are requesting features, let’s work on those features. And you can tell from the outside, to your question, you can tell from the outside whether this is happening or not, right? There’s a lot of signals, right? Has the website been updated? Is there a recent release?

Ben Langmead: Are they, are there, is there anything going on the issue tracker, right? And I say this and I think, I hope people don’t check all my issue trackers on all my tools ’cause I’m not proud of what you’ll find in all cases. But we have ways of from the outside interrogating to figure out if this is having a long and healthy life in this lab, or if maybe the lab has turned their attention elsewhere and maybe this thing isn’t too usable.

Ben Langmead: Documentation is another very important thing. Has the team done something more like a reference implementation of an algorithm and then moved on? Or did they think through with empathy for the user? What will the user actually do and be confused by, and can we break it down into subcommands that the user understands what they are and how to sequence them in order to get to their, to their goal?

Grant Belgard: Your lab has always been really good at that, by the way.

Ben Langmead: I appreciate that and it’s ongoing communication, right? And sometimes it has to be face-to-face, right? There’s definitely tools where I think everything is going a certain way. I’m interacting with the issue tracker. I’m having weekly meetings with my software engineer.

Ben Langmead: Then I go out and give a talk and have a meeting with somebody, and I realize, oh, this person thinks about what they want from this tool in a totally different way than I expected. And then, next meeting I have with my software engineer, I’ll relay this and I’ll say “maybe this means that we should provide this additional command, or maybe this means that we need to take it in this direction.”

Ben Langmead: So it’s ever-evolving and it definitely requires nice high bandwidth communication with the community, which doesn’t always mean it can be limited to things like GitHub issues or emails. It sometimes has to be go out in the community and ask people questions and figure out what they’re confused about or what they’re worried about.

Ben Langmead: But yeah, we all know the difference and my take would just be it’s okay for software to go out there and not get traction and gently, sink back into the earth over time. That’s perfectly okay. I think that is important for a healthy research ecosystem.

Ben Langmead: PhD students need to have projects, and nobody knows which are the ones that are gonna get traction. And so if everything works out just right, then the ones that should get traction do, and the people who made it have the time to pay attention and do the documentation and have the empathy to make it truly easy to use and also commit to it long term and go out and talk to people about it when they’re at conferences and talks and such.

Grant Belgard: What do you decide when software is ready for other people to use?

Ben Langmead: I think it is true that some of what we put out is more of the form of a reference implementation. So in other words, that software is there because the idea is good, and in order to show that it was good, we had to implement some software. And that’s not always software where I’m gonna expect a white coat-wearing biologist to come along and download it and use it.

Ben Langmead: And again, that’s okay. There should be a notion of a reference implementation. And in the day of AI coding and AI rewrites, it could even be that reference implementations are the main thing that people like me put out. Because if it’s gonna get traction, perhaps it’ll get traction through an AI rewrite of my reference code rather than my code per se.

Ben Langmead: I like to think… first of all, when you write the documentation, that’s a pretty good stress test of whether you really did think through how users are gonna use it, right? ‘Cause if your documentation is reiterating your paper, that’s probably not a good sign.

Ben Langmead: But if your documentation is saying, “If you have this problem, then run this command,” and you thought through what are the problems the user might have, then that’s better. So it becomes like a question of what are the nouns and the verbs? Do they map onto things that biologists actually already know?

Ben Langmead: Can we use standard file formats like when a biologist comes along to interact, are we gonna be introducing them to new file formats? That often is a point of friction. Or are we telling them, “Yeah, you can use the stuff you’re very used to using.” You can think of it as a converter if you want to.

Ben Langmead: It’s true, we put a lot of effort into the algorithm, but sometimes things really do fit into a pipeline and play a role in a pipeline, and we need to allow the user to think in terms of getting from point A to point B in their pipeline and not get too hung up on what’s our algorithm. So yeah, I think that if you can think in terms of the user’s problems and nouns and verbs, and if your documentation is written like that as opposed to, “Here’s everything we did,” ” here’s all the functions we implemented,” then I think you’re well on your way.

Grant Belgard: What first pulled you towards computer science?

Ben Langmead: Gosh, computer science I was into from roughly high school. Maybe even middle school. So I was at a middle school where they taught us Basic, the programming language, and I could tell that it was fun, and I could tell that I was good at it. And so when high school came along, I took all the computer science courses I could.

Ben Langmead: Also, my brother, who’s eight years older, he was taking computer science classes in college, and I was asking him to send me his homework so that I could work on them too, and, I was trying to learn C programming. And so by the time I got to college, I was very sold on the idea that I really like the ability to tell the computer what to do, and that it truly does what I told it to do.

Ben Langmead: And then when I got to college, I knew right away I wanted to major in computer science, and I took my computer science courses. And I think what I further fell in love with was computational thinking and thinking about software and hardware. You mentioned high-performance computing as one of my emphases at the top and that’s true.

Ben Langmead: In fact, before I went to graduate school, I basically only did high-performance computing. And my favorite thing about that is like un-un-black-boxifying the computer, right? Being able to think of the different pieces of the computer so that it’s not a mystery. “I ran this, I was sure it was gonna be fast, but it was slow.

Ben Langmead: Why did that happen?” Trying to demystify these things and say, “No, I get it. I see how that mapped in an unfortunate… It was a good idea, but it mapped in an unfortunate way onto the way the hardware works. I need to think about the way the hardware works and come up with a new idea.” So I was, I’d say, sold from an early age on the idea that programming is fun and it’s just a fantastic hobby, and I, and if I could do it as part of my job, that would be great.

Ben Langmead: And then in college, I think I fell even more in love with making the computer something other than a black box. And then the next step for me was graduate school, which is where I truly got to commune with the idea that computer science can be used for basic science, which is not necessarily something that computer scientists ever learn, right?

Ben Langmead: Like you can go through an entire computer science undergraduate curriculum and think the whole time that computer science equals big tech. Computer science equals Google, Microsoft, Facebook and of course I’m gonna move to San Francisco when I’m done, right? But going to grad school, at the time I did especially, taught me no.

Ben Langmead: People who are computer scientists can work shoulder to shoulder with basic scientists, and that was the next step in my evolution. And that piece I think is… like my story is a very common story, I’d say right up until grad school because that final piece of connecting computer science with another science is not something we do so well in our curricula.

Grant Belgard: Was there a moment when you realized that genomics had problems you wanted to spend your career on?

Ben Langmead: Totally. Yeah, when I got to graduate school, I was working with one professor at the time at University of Maryland, Bill Pugh, and I was not working on genomics. But at the orientation events, I went to an orientation event where Stephen Salzberg was speaking, and he was describing next-gen sequencing which at that time was brand new.

Ben Langmead: And like I say, he had probably only seen one or two, Solexa data files at that point. And he was talking about what a big deal it is that it’s so inexpensive now, and what a big problem it is that we can’t– we don’t think we currently have the tools to analyze these. And I was so compelled by that, that I went up afterwards and I asked him “look, my background is this and this.

Ben Langmead: Like I, I kinda do high-performance computing, and I’m very much a computer scientist, and I don’t know, last biology class I took was freshman year of high school. Is this an area where I can work?” And he said, “Absolutely.” He was like, “We need people like you.” And he told me to talk to two of his students at the time, Mike Schatz and Cole Trapnell.

Ben Langmead: And luckily for me, Mike and Cole were giving a seminar shortly after that on something they’d been working on called MUMmerGPU, which was a whole genome alignment program that used CUDA and GPUs. This was in, the year 2007 or 2008, so it was very early for that kind of thing. And when I went and saw their talk, they gave a tag team talk, I was at that point completely sold, right?

Ben Langmead: I was– I wanted to be like them. I would do anything to be like these guys. And so I started to talk more to Stephen and also to Mihai Pop, who was my master’s advisor along with Stephen Salzberg, and try to figure out, okay what are the handholds we can get on this alignment problem?

Ben Langmead: And by the first summer after I started grad school, I was fully only working in that area.

Grant Belgard: For a student who likes both biology and computing, how should they choose what to learn deeply?

Ben Langmead: That’s a great question and I can only step on toes by answering your question, so I’m just gonna go ahead and do it. So what I always tell students who approach me and ask that is, I think the computational toolbox is the one to learn earlier. If you’re gonna learn– If you’re gonna sequence these things I’m gonna do one and then I’m gonna do the other, or I’m gonna do one and I’m gonna apply it to the other.

Ben Langmead: I think the thing that’s important to get early is the computational thinking. It is possible to then pick up more of the biological thinking. Okay? That’s my opinion, right? This is not what I would say necessarily to my biology colleague, but this is what I say to students. So I feel like if you’re, if you have that computer science training and toolbox, there’s not really a door that’s closed to you.

Ben Langmead: I don’t think there’s too many professional opportunities where they’ll say, “But you’re a computer scientist. We’re, we do XYZ.” As long as you can demonstrate that you are, that you have the computer science, but you know how to apply it and you know how biological people think, I think it’s perfectly fine to have that computer science degree.

Ben Langmead: So I’m often talking to students who are trying to weigh, “Do I wanna major in computer science and then minor in something?” Or maybe they’re looking at grad school and they’re saying, “Do I wanna do a computer science PhD program or maybe a, a biomedicine or a computational biology or some other…”

Ben Langmead: I usually tell them, “I think that computational toolbox is the crucial one. It’s important to learn it early, and so it’s good to put your academic emphasis on that and grow into the other areas.” I think, like one of my experiences early in my career was before I got my PhD, I worked as a research associate at Johns Hopkins in biostatistics with Rafat Irizarry, a biostatistician, and he was collaborating with a epigenomicist, Andy Feinberg.

Ben Langmead: And I would go to Andy Feinberg’s lab meetings, and at first, I understood basically none of what I heard. It was like being immersed in a foreign language. But if you keep going, and maybe also if the students are willing to answer your questions after the lab meeting if you didn’t understand something, you start to see s- things, certain things repeating, you start to see certain themes, and then you do pick it up after a while.

Ben Langmead: And so I feel like that is a possible thing to do, whereas I’m not sure quite how possible that is with computer science. You have to bang your head against, “My program is not compiling. My program is not working. My algorithm is taking 10 times longer than I thought it would.” Those are things you have to commit to long periods of time of banging your head against it before you can get to the end.

Ben Langmead: So I usually advise that the toolbox to get early and to have represented primarily in your credentials is the computer science one. Now, I’m obviously biased and I always tell students that. So listeners should know that was a biased opinion.

Grant Belgard: To close this out what advice do you find yourself giving again and again?

Ben Langmead: Okay, so I interact with a lot of students at all levels, undergrad and graduate students. And so I think when I’m talking to undergrads, it’s usually because they’re trying to figure out how to map out their college experience and what do you do? And they’re trying to figure out, should I cram in as many classes as possible, or should I triple major?

Ben Langmead: And what I’m usually telling them is something like it’s actually quite difficult to predict what it is you did as an undergrad that’s gonna unlock that next door in your career. It might seem like you want to take as many classes as possible in the thing you’re most interested in, but my story is when I finished undergrad, I applied to several jobs, and I ended up getting a job at a small consulting company called Reservoir Labs.

Ben Langmead: And after working there for a couple years, I was talking to my boss, and he revealed, “Basically, the reason we hired you, the reason we picked your resume out of the pile, is I really liked your writing sample.” And this was, like, a essay about Japanese history from one of my East Asian studies classes, right?

Ben Langmead: So could I have predicted that my first job out of undergrad as a computer science major would have been based on my writing sample? No. And so I think that as important as it is to craft the right undergrad curriculum, it’s also important to explore and make sure that you’re developing that wider skill set: communication, writing presentation skills, as well as whatever it is you’re doing.

Ben Langmead: And then for grad students, I would say one of my, I have lots of advice for grad students, what does it mean to be a PhD student? But I think one of the messages that’s hard to take but important is failure is an option, right? This is science.

Ben Langmead: Your most beloved hypothesis the underlying science doesn’t care how much you love it, right? It’s you have to go with what the data is telling you. And unfortunately, that means that you can’t point at the calendar and say, “This thesis chapter is gonna be done by this date,” or at least not not without being frustrated at the end of the day.

Ben Langmead: So I think people who go to grad school are committing themselves to allowing science and the data to tell them when they’re done and when they’re not done, and that’s a tough commitment, but it’s the right one.

Grant Belgard: Ben, it’s been a great conversation. Thank you so much for joining us

Ben Langmead: Thank you. Pleasure to be here

Transcript of Episode 87: Elliott Margulies

Disclaimer: Transcripts are automated and may contain errors.

Grant Belgard: Welcome to the Bioinformatics CRO Podcast. Grant Belgard, joined… i’m delighted to welcome Elliot Margulies director of bioinformatics at Billion to One, a molecular diagnostics company applying quantitative approaches to prenatal screening and liquid biopsy. spent his career at the intersection of genome technology, computation, and clinical translation, including earlier work at NHGRI and Illumina. In this episode, we’ll talk about what he’s working on now, how his career developed, and what advice he has for scientists and engineers who want to build bioinformatics that matters in the real world. Elliot, welcome to the show.

Elliott Margulies: Thanks, Grant. It’s really great to be here. Great to see you again too

Grant Belgard: uh, The pleasure is mine. for listeners meeting you for the first time, what are you working on now and what problems are most alive for you day to day?

Elliott Margulies: Well, I’m like you said, at a molecular diagnostics company, and so day to day uh, my team focuses on the processing of uh, data uh, for prenatal testing. And we focus on everything from kind of the infrastructure uh, and analysis of kind of primary sequence data, making sure the instrumentation is running correctly, our QCs are, are appropriate, all the way through to building and maintaining curation and interpretation systems and building out reporting platforms.

Elliott Margulies: We’re, we’re kind of the glue across all the different parts of the organization that has to touch a sample from when it’s, uh, acquired to when it’s reported out, uh, whether that be the lab, our genetic counselors, lab directors, uh, even some software teams who are managing, uh, interfaces, uh, and portals that, uh, our physicians and patients interact with.

Grant Belgard: Would you explain the core bioinformatics challenge in your current work to a genomics audience that’s new to this particular area?

Elliott Margulies: Well, we very much focus on actionable, informative, places in the genome. Uh, This is actually, uh, a kind of a, a migration that I’ve made in my own career, uh, from, you know, figuring out algorithms and methods to interpret and analyze whole genomes, uh, bringing that to scale and thinking about how healthcare systems might adopt that, to now being at a company that focuses very much on very specific parts of the genome, uh where, um, we are able to determine with high confidence, uh, whether risk variants or other de-deleterious variants are present.

Elliott Margulies: Uh, and we, and, uh, we can do so both by looking at, um, the DNA from mom, um, but we can also look at the cell-free DNA, uh, that represents the growing fetus inside of mom, uh, and determine the probability that, uh, the baby is carrying zero, one or two, uh deleterious alleles.

Grant Belgard: Can you make technical work understandable across teams with very different vocabularies?

Elliott Margulies: Oftentimes we remind ourselves that everybody is in a silo to a certain extent with their own lingo, and we make it, uh, comfortable for people to pause and say, “Wait, I don’t understand you,” or, “What was that acronym you used?” And, uh, we really have a fun culture of kind of acknowledging that everybody is an expert in their own field, but when th-those interfaces start to talk to each other, you gotta slow down and, and not be afraid to say, “I don’t understand.”

Elliott Margulies: It’s actually a really strong part of our culture, um, at this company right now where that understanding of constant learning and explaining or, um, feeling comfortable saying, “I don’t understand,” or feeling comfortable saying, “Ah, I think I may have screwed that up,” uh, you know, “Let me fix this,” or, “Let me learn what, what happened with the process that we can improve upon.”

Elliott Margulies: But it all starts with, uh, clear communication and feeling comfortable, uh, at those interfaces and slowing, slowing the conversations down, up-leveling, using analogies

Grant Belgard: What do you wish more people understood about the bioinformatics behind clinical testing?

Elliott Margulies: We’re, incredibly rigorous, right? So when I think about the reproducibility, and the confidence we have to have that all of our pipelines are working, we’re constantly thinking about how can we cross and double and triple check that everything is working correctly, that we’re giving the, the right answers to the right people all of the time.

Elliott Margulies: And so, the– running a pipeline is easy, but running a pipeline at scale with a level of confidence that every patient is getting the right answer, as often as possible, and that we have ways of… ‘Cause oftentimes it’s the exceptions that we spend the most time on, right? So, fortunately in a prenatal testing situation, most samples are negative, and that’s g- that’s a good thing.

Elliott Margulies: or I guess another way of thinking about it is that most of the time, sequencing runs and analyses work out just fine. But, we spend the majority of our time on those, those small exceptions, whether it’s a complicated case, a sequencing run that didn’t go as well as we thought, some other issue that’s kind of coming into the pipeline.

Elliott Margulies: We spend the majority of our time, right, thinking about, you know, those small number of samples that are not within the, the norm.

Grant Belgard: When you encounter those, edge cases, are they typically treated, as, one-offs, or, do you typically make changes to the pipeline or process or something on the basis of those? Or are they just so rare and weird that they, typically don’t trigger kind of

Elliott Margulies: No, I’m actually gonna quote, a team member of mine, his name is James Hart. we were recently talking about, building new pipelines for future products. And, I s- I, we were talking, and I had said something to the effect of, you know, “We need to make sure we can handle the edge cases.” and he looks at me and he says, as he’s, has a great, mindset when it comes to building the- these types of, infrastructure.

Elliott Margulies: He says, “Elliot, there are no edge cases. There are only workflows that are run rarely.” And it was a, it was a nice epiphany, I thought, where, And it was, I was like, “Oh, we’ve got the, the, the right person thinking about this.” But it is true in the, in the sense that for some of the legacy systems that we manage, when there are edge cases or one-offs, we see them with some level of repetition.

Elliott Margulies: And particularly as things scale, it’s important for us to build systems that can more handle those types of situations in, in a more standard way.

Grant Belgard: So I guess on that note, how do you decide whether a problem needs a better assay, a better model, better data, a better process?

Elliott Margulies: the, you know, there are many types of rules and kind of prioritization, right? So you, you have, you know, a finite set of really smart people on your team, who are balancing both kind of production responsibilities as well as building new products, as well as automating and scaling kind of existing infrastructure.

Elliott Margulies: And we balance our time across these things, and oftentimes we l- we look for, you know, the proverbial low-hanging fruit or take that 80/20 rule. You know, where, where can I spend a small amount of my time and, and have that be a large impact? And oftentimes we think in that, in that way when we prioritize work, particularly over the short term.

Elliott Margulies: And so we try and balance the short term, where can we do a small amount of effort that’s gonna have a big gain? and a- another metaphor sometimes we talk about is those types of small efforts that have large gains start to get a flywheel going of, more automation, and that more a- additional automation frees up even more time to then automate even further.

Elliott Margulies: And before you know it, you’re starting to, to spend time thinking about the larger system and the longer term growth that, that is necessary.

Grant Belgard: Yeah. How do you think about uncertainty when a computational result may affect a clinical report?

Elliott Margulies: We’re very much a first principles-based company, and so a lot of the analysis that we do, uh, is rooted in statistics, uh, and probability and, and quantifying, um, you know, where we, we have the ability to essentially count molecules, um, at a– that are at a very low level. Uh, but we can do so very– in a very quantitative way.

Elliott Margulies: And so oftentimes we are, uh, we, we put probabilities around, um, all of our calls. And, uh, we are constantly looking at, um, outcomes data. Uh, so we have a fantastic outcomes team that, uh, of course it takes, you know, nine months or so to, uh, to get the, uh, the outcomes. But, uh, over time, these, these things build up, and we scrutinize our false positives, our false negatives.

Elliott Margulies: Um, they’re very, very small, and they’re, they’re rare. Um, but this is a screening test that we are looking at, and, uh, those, uh, can lead to ways in which we can think about systematically continuing and to improve the assay.

Grant Belgard: Makes a quality metric useful rather than ornamental

Elliott Margulies: if it drives at, at something that is measurable and changeable. and again, that’s, that’s very much in our kind of ethos is this concept of being guided by first principles. We often want– oftentimes want to measure something directly versus indirectly. I think as bioinformaticians, it’s easy to kind of, pull together some sort of, you know, convoluted…

Elliott Margulies: I’ll call it convoluted correlation. but if you can’t detangle, you know, what are the values and metrics going into why you’re seeing this correlation, it becomes abstract and, you don’t know what you’re missing or what you’re not seeing. And so oftentimes we try and peel away the layers of the onion so that the analyses we’re looking at are rooted in first principles.

Elliott Margulies: And we’ve seen successes, in doing that. I can think of a recent example where we were doing an analysis and, you know, we could draw, you know, a weird square around, you know, certain data points and kind of increase our sensitivity, and, you know, with a minimal, you know, increase to no-call rate.

Elliott Margulies: But, at the end of the day, we actually found that, doing a more principled Z-score approach, was much cleaner. and we, we understood what was going into that Z-score, in a, in a much better way. And so, so that’s, that’s constantly on our mind as we, as we’re doing these analyses.

Grant Belgard: Where do you see AI being genuinely useful in molecular diagnostics?

Elliott Margulies: Oh, fun question that I know is, is, p-potentially talked about all the time. But even over the last six months, we are seeing bioinformaticians turn into full stack software developers, but with this scientific level of expertise where they deeply understand the workflow and the systems.

Elliott Margulies: And so they’re in some ways hyper-enabled, to move things forward. We’ve gone from, you know, building out prototypes that involve, you know, tab-delimited files and filtering them through Excel spreadsheets to now, you know, on the first try, having a customized user interface, you know, that, is designed for the particular workflow that we’re, we’re doing.

Elliott Margulies: and, that’s all empowered because we’re historically not front-end developers, but we’re now empowered to be able to be front-end developers, with the assistance of, of AI. And so I’ve seen it– I’ve seen the, the aggressive adoption, aggressive but thoughtful adoption, both as a software development tool, and also as a data analysis tool.

Elliott Margulies: So you can quickly advance a variety of different analyses, with AI assistance. We, you know, I think, again, we’re hyper-enabled because we’re already on the command line. We’re already accessing our databases and our infrastructure and our systems. So the access to the data is not a problem.

Elliott Margulies: The envisioning of what we want to do is not a problem. And so we can very well prompt and describe to AI systems what we’re looking to do. And I’ve, I’ve watched, both on the software development side and on the data analysis side, the speed with which we can move is just incredible. You almost have to pause because a human being can’t absorb it all as fast as we’re able to generate it all, right?

Elliott Margulies: And, and that’s actually, we recently had a hackathon as a team, and one of the comments was, you know, “Yeah, I built this thing so fast, but I, I still need to like pause and like look under the hood and really fully understand w- you know, what was going on here.” And, and, and, and that, that disconnect is, to the actual line-by-line coding is what everybody is starting to get used to and try and, and, and work through.

Grant Belgard: What changes when bioinformatics moves from method development into high throughput clinical operations?

Elliott Margulies: There’s, there’s a little bit more of a rigor, uh, associated with change control, um, ensuring the quality of the output is what you expect it to be. In many ways, you know, we strive for things to be as perfect as possible, but more importantly, knowing where, where it falls over and documenting that and under– and, and having an appreciation for that is, uh, in some ways more important than getting the algorithm perfect.

Elliott Margulies: We don’t– We, we never wanna be blindsided by an error or an issue that we didn’t realize our systems weren’t performing as well as they should. And, um, so we’re on– constantly on the lookout for, for that from a clinical ca- uh, standpoint, particularly also as the scale come, comes about. Um, but we have, you know, fantastic team members who think through a problem, and anytime we’re able to, um, make an improvement, one of the exciting things is that we’re sitting on, um, a treasure trove of historical data that we’re able to understand and with, with outcomes information.

Elliott Margulies: If we make an improvement, how well does it perform? How did it– How does it, uh, compare to the way our, our, you know, previous version was with, with the system? And so we, we build automation and systems that are able to essentially do an AB comparison, right, off of, you know, tens of thousands of samples, if, if not more, um, uh, to, to be able to have that rigor whenever we push a, a new change into the system.

Grant Belgard: How do you know when a pipeline’s ready to leave R&D and enter routine use?

Elliott Margulies: I mean, the, the formal way is through a, a, an actual validation. We set up a, you know, a multi-arm trial, that kind of tests both kind of individual measurements as well, you know, and then it grows from there to reproducibility to final kind of validation outcomes.

Elliott Margulies: At, at the end of the day, we’re, we’re making a, a variant call or a, an assertion of risk, and we have to have a known truth set, and we have to, show reproducibility. and then we also try and show, you know, almost like forcibly show limits of detection. Where is there flexibility in the system where even if it doesn’t, if it’s not perfect, it’s a little bit noisier data, we’re still getting the right answer.

Elliott Margulies: Like understanding that limit of detection is, is really important. But it all happens through, you know, we, we follow very standard, processes by which we verify and validate our, our, algorithms.

Grant Belgard: Now to, shift gears and talk about your career. what, early experiences most shaped the scientist you became?

Elliott Margulies: I can point back to a lunch I had as a grad student where Francis Collins was, a professor of genetics at, University of Michigan. He was now at the NIH, but he had still come back on occasion through his relationships with the University of Michigan, at the medical school, giving lectures on occasion.

Elliott Margulies: And hearing him talk about his day moving the Human Genome Project forward, is still a core memory of mine as a kind of a young, graduate student. and that planted a seed in my head about how you can take, a doctorate degree and not just be a, an academic professor writing research grants, for the rest of your career.

Elliott Margulies: Kind of the desire to collaborate on a scale that was beyond what one person can do, it started to really kind of be a seed that was, you know, sprouting in my brain. And, and then moving to the NIH for my postdoc, with Eric Green, who I know you’ve recently had on this, this podcast as well, really opened my mind to that type of at the time it was called big science, right?

Elliott Margulies: Where you could, generate data at scale, you could collaborate at scale, and it was a ton of fun. I’ve, I’ve, I’ve constantly found myself moving in, in those directions where the learning from a diverse group of scientists, is feels much more comfortable to me than working on an individual a-assay or experiment, or project o-on my own.

Elliott Margulies: I have those little pet projects here and there, but at the end of the day, it’s watching the, the, the group of people accomplish something that was not possible as individuals. You, you know, that’s, that’s been a ton of fun for, for my career.

Grant Belgard: Can you tell us about, some of the later major inflection points in your career?

Elliott Margulies: Sure. So I, I sometimes

Elliott Margulies: think about, my career, at least right now, as, having kind of three major phases, I guess. both– So I started as a faculty member, at the National Human Genome Research Institute, where I was doing more basic science research, but in this kind of largely collaborative way, trying to make an impact on the world with, this transition t-from, kind of population scale, you know, low coverage sequencing to find variation in, in the human genome to being able to sequence individual genomes.

Elliott Margulies: And how do you build the methods that can, accurately call variants in individual genomes? I then, kind of the second phase of my career is when I moved to Illumina, working in this advanced medical research group, with David Bentley, and his team, where we really thought about both kind of balancing, what is possible, and from a clinical medical perspective at scale, and, balancing that with kind of the drive of the next generation sequencing that, Illumina at the time was, was driving forward.

Elliott Margulies: and it was a fun balance, because we were trying to provide y-you know, utility to the sequence data that was being generated, and added value and thinking about, I was able in that environment to think five to ten years out, you know, how can we build collaborations to show what’s possible about bringing, you know, high throughput, whole genome sequencing to a medical system?

Elliott Margulies: You know, at the time, you know, we had recently, started the hundred thousand genomes project in, in England as, as a kind of a f-first initiative of bringing genomes into a clinical context. and then this third phase of my career, which I’m in right now, which to me, there’s a bit of, irony in the fact that for much of my career, I had spent focusing on whole genome methods for, you know, sequencing and interpreting, genomes, to now sequencing super small parts of the genome that are incredibly actionable and meaningful and doing that at, at scale, and, being so close to patient care.

Elliott Margulies: this is something that is, is incredibly meaningful to me and my team. And I don’t think it’s often you get a bioinformatics group like we have here that when we find a high-risk result, it weighs on us, and we know that we’re now supporting genetic counselors who are gonna be giving this information to families at a very, you know, that are, the, the, at a very challenging time in their lives.

Elliott Margulies: But they’re going to be better for it because they have this information. We really believe that having that information is, is empowered, empowering to families. but we are, as a bioinformatics group, so close to those decisions, and supporting teams who are delivering this information. It really is remarkable and, in other parts of my career, even though I was building tools that were interpreting genomes or building methods that could identify variants in rare diseases, I was never this close to patient care, and it’s very rewarding, to be able to, to support an organization, like this.

Grant Belgard: How has your view of sequencing changed as the field has matured?

Elliott Margulies: Well, I think we’ve reached a point where, uh, sequencing is a, a commodity. Um, I mean, I r- I remember a time when it was very novel to be able to do massive short-read sequencing. Um, that is now a commodity, and it is just one part of a toolkit that enables clinical adoption. Um, and something that I’ve learned a lot from this most recent part of my career is how much do you have to think about holistically, a, a sound business model that is reimbursable and you’re building a system that, uh, can operate at scale, uh, within certain cost measures and where sequencing is just one part of, uh, of that cost.

Elliott Margulies: Um, but thinking about how to interact with insurance companies, with physicians, um, with patients is all part of the process now. So it’s, it’s become… It used to be a, I guess, to, to answer your question, it used to be a major part of everything that I thought of, uh, when– as a bioinformatician, and now it’s just one part of, of many that are equally, if not more important, because we’ve essentially succeeded in making DNA sequencing a commodity.

Grant Belgard: What have different work environments taught you about how science gets done?

Elliott Margulies: It’s all about the people and the relationships. I’ve been around smart people I think all throughout my career, whether I was in grad school, as a postdoc, um, at Illumina, here at Billion to One, and the times where we were successful and impactful were the times when we could get large groups of people working well together and feeling comfortable and open to uh, making decisions under difficult situations, not being afraid to call out when there are issues or concerns or timeline risks or technical risks.

Elliott Margulies: Um and I can point to the exact opposite happening, where, um, sometimes great ideas didn’t succeed because people were afraid to say, “I don’t think this is gonna work,” or, um, “I don’t know if this other person agrees with this, this plan forward,” right? And you know, everybody can be well-intended, but, um, if you don’t have that open, uh, way of communicating, uh, and being able to, um, talk about the, the more challenging things in a calm, open way, um, uh, oftentimes you can’t see these, these great wins and these great successes.

Elliott Margulies: Uh, and I found a company that truly values all of those, those parts of the, the scientific process and the development process. So I really feel supported, both w- my kind of senior colleagues, as well as the team members that, uh, I’m mentoring, um, to be– to foster this type of environment.

Grant Belgard: What’s something you’ve learned the hard way?

Elliott Margulies: I think finding an organization and a mission that is aligned with your own career goals and growth is important. I can think of a time in my career where I thought I was well-aligned with where I was trying to move things forward, but it ended up not being the same as the organization that I ended up being in.

Elliott Margulies: And for a period of time, I really struggled. Not because somebody was, was trying to set me up to fail, but because my interests, uh, and where I thought the organization wanted to go was different than where the people who were leading the organization wanted the organization to go. And, uh, it took me a while to realize that disconnect.

Elliott Margulies: And once I kind of realized that and I was fortunate enough to be able to kind of move into a different organization that was well-aligned with my goals um, my career started moving forward again. And I think that’s, um, often-oftentimes, right, nobody’s ill intent, but, uh, when you get that mismatch of, like, where you want to go, what your passions are, and if that’s not aligned with the organization that you’re in, um, that can be a rough, rough time.

Elliott Margulies: But to your point, I can also look back on that and say I learned a lot during that time that, uh, I wasn’t necessarily move-moving my career forward. Um, in hindsight, it has helped me move my career forward because there was a lot I learned during those periods of time.

Grant Belgard: Related to that, have mentors taught you that still shows up in your leadership style?

Elliott Margulies: I can point to very specific things that I do that are relics, so to speak, of things that I learned because mentors have done them to me. And some of that is fostering that deep collaboration. So Eric– I can think about Eric Green, uh, working in his lab. He just opened doors to, for me to collaborate with people that I didn’t, uh, know I could collaborate with, and I learned so much about that.

Elliott Margulies: And I enjoy when I can bring people of kind of from different parts of the company together and start to see them, their minds just go like, “Oh, wait, I’m thinking about this problem too.” And, and they’re thinking about it from a different angle, and you kind of see that, that collaboration take place. I can think about, you know, as I move my team forward, um, and I constantly am trying to think about how are they…

Elliott Margulies: Are, are they happy? Are they growing? Are they doing what they want to do? Uh, w-when we get a chance to be together, recently we had our, our teams off-site, um, we all shared a success and a challenge. And giving everybody the space and the comfort to share a challenging moment that they might be going through with the, with the entire team in a way that is very healthy and productive, and they don’t feel like they’re going to be shamed or, um, or, or disadvantaged because they shared something they’re struggling with.

Elliott Margulies: To create that environment, it becomes very empowering, and that’s something that David Bentley taught me. Uh, you know, so I can think about, you know, that, that type of interaction of leading a team that’s very thoughtful, um, and constantly trying to grow and these are things that, that, uh, definitely come from great mentors that I’ve had in the past.

Elliott Margulies: On a lighter anecdotal note, uh, music is really important to me. Francis Collins is somebody who always brings music to a lot of the science that he does and the way in which we kind of bring teams and communities together. Um, I started my off-site by singing a song to everybody and having them sing with me.

Elliott Margulies: Um, these are things that I, I often do that are fun to think back on how wonderful leaders and mentors have kind of touched my own life and, and career and how I’m trying to give it forward to the, the rest of the team that, that I now lead.

Grant Belgard: What do you look for when hiring bioinformaticians who will thrive in clinical diagnostics?

Elliott Margulies: So we hire people at different stages of their career. I’ve been very impressed with the tenacity and thoughtfulness of the right person who is kind of just coming out of an undergrad or a master’s degree. The work that we do can, be very impactful and has to be done in– with, with a lot of thought, because these are patients who may be getting a high-risk result.

Elliott Margulies: And it takes a certain kind of individual who has a level of maturity, to, understand that what they’re building has to be right as best as they can do it, where you have to own up to a mistake, in a, in a way to– that can be productive, where both you learn and we learn how to improve the system.

Elliott Margulies: so I think finding a smart individual who can code and run bioinformatics pipelines is kind of table stakes. It’s really looking at an individual, what are they passionate about? How do they wanna make an impact in the world? How do they wanna grow their career? Are those passions aligned with being in a company where, we are managing patient samples?

Elliott Margulies: being in a company where we’re doing science that is leading to new products, where we’re going to be balancing an intense period of time of scaling and b-and building new products and trying to automate more, so that multitasking ability. But at the same time, tr-tr-uh, trying to ensure that people can focus and not get overwhelmed, with everything that’s going on.

Elliott Margulies: So these are, these are kind of traits we try and, and look for. you know, these are insights that, that go beyond just coding. But, you know, can you, can you analyze a data set and, you know, describe what you’re doing while you’re coding and, you know, present the, the results to people, in a meaningful way that are, you know, getting back to one of your earlier questions, you know, that the cross-collaboration, is key.

Grant Belgard: What advice would you give to computational biologists who want their work to have clinical impact?

Elliott Margulies: Well, definitely align yourselves with, a company or a lab that has a clinical focus. I think more and more, we’re seeing the ability to translate genomics into actual clinical results. And, it’s not, you know, to something we were talking about earlier, it’s not just a matter of sequencing somebody’s DNA and doing some variant calling anymore, right?

Elliott Margulies: I mean, there’s a whole system and an organization of how do you interact with patients? How do you interact with healthcare systems? How do you interact with reimbursement? and being able to… Something that I don’t think I fully appreciated till I joined Billion to One was that this organization has that all, and this is why I think I’m feeling empowered to succeed bec- in a clinical context.

Elliott Margulies: Whereas I thought I was doing clinically impactful work, and I probably was, right? But it was more futuristic research. But what I’m doing right now is, is impacting patients today, and, that’s very, very meaningful. So looking for an organization that is thinking holistically about all the different aspects of molecular diagnostics, not just the molecular part of molecular diagnostics.

Grant Belgard: What should trainees spend more time learning than they currently do?

Elliott Margulies: I think understanding a lot of the science and engineering principles is b- is going to be incredibly important because coding as we know it today is starting to go by the wayside. it is somewhat of an identity crisis for bioinformaticians. but it means that understanding how software systems are architected, that’s not gonna go away, right?

Elliott Margulies: you still need to guide an AI to build the right tool in the right way. and making those decisions, requires knowledge of how to build scalable software systems. There’s a kind of an architectural engineering component that I think is gonna become more and more important, as well as the science, of things, right?

Elliott Margulies: B- being, you know, being able to, continuously learn and adapt to the science of what you’re doing, will be essential, in, in moving things forward. We’re, we’re no longer just people who plumb and build pipelines. that’s not what we do anymore.

Grant Belgard: If you could go back and give advice to your younger self, what would, would it be?

Elliott Margulies: don’t underestimate serendipity

Grant Belgard: And, to, to wrap things up, are you most optimistic about in genomic medicine over the next decade?

Elliott Margulies: Finding deleterious pathogenic variants and, medicines that can overcome those deleterious mutations, that is an incredible change where we, you know, first we had to develop the tech to identify mutations. Now we can identify mutations. And soon, not only can we identify these mutations, but we can proactively provide pharmacogenomic, you know, information, and, and drugs that can mitigate, those mutations.

Elliott Margulies: That’s just an incredible place to be. We’re seeing this in the prenatal space now, start– just starting to emerge with a couple of key examples. but that’s, that’s a wonderful place to be because now you can kind of complete the cycle of not just delivering challenging news, but, ways in which you can mitigate, those, those mutations.

Grant Belgard: Have any, any closing advice for, career scientists and bioinformaticians listening to the show?

Elliott Margulies: Focus on what makes you happy, uh, you know, and, and curious. Uh, stay curious. I think, um, it’s important to realize that what you learn today is not necessarily what you’ll do in five or 10 years. But maintaining that ability to constantly learn and adapt, uh, is, is a, is a fun place to be. And again, you know, lean into the serendipity part.

Elliott Margulies: You, you never know, uh, when, uh, you’re gonna work on something and the– this particular project is gonna be the extra meaningful thing. Uh, you didn’t plan for it, but it’s turns out that the next three months are, uh, uh, an, gonna be an intense moment where you might learn an extraordinary amount in that short period of time that you didn’t plan for, and you’ll make a, a disproportionate impact by investing in, in those, those periods of time.

Grant Belgard: Elliot, thank you for joining us.

Elliott Margulies: Oh, it’s been my pleasure. Thanks for having me, Grant

Transcript of Episode 86: Deniz Kavi

Disclaimer: Transcripts are automated and may contain errors.

Grant Belgard: Welcome to the Bioinformatics CRO Podcast. Today, I’m joined by Deniz Kavi, co-founder and CEO of Tamarind Bio. Tamarind is building software that helps scientists access computational biology tools at scale through a web platform and API across areas like protein design, structure prediction, and docking.

Grant Belgard: We’ll talk about what Deniz is building now, how he got there, and what advice he has for people trying to build at the intersection of biology software and AI. Welcome.

Deniz Kavi: Thanks for having me

Grant Belgard: Thanks for coming on. So tell us about Tamarind

Deniz Kavi: Yeah. Tamarind as you described, basically a aggregator of all the leading molecular design tools. So things like AlphaFold and RF diffusion and many others, we provide in a single place for all scientists to use, basically. So we try to do the work, like the boring, annoying, undifferentiated work is what we try to do our best in.

Deniz Kavi: So yeah, scaling up these tools, applying them to real problems, connecting them into multiple stage pipelines, so on. I got into this problem basically doing this as a human for my colleagues at Stanford. So in my undergrad lab, my job was to get an email, run AlphaFold 10,000 times, email it back to one of my colleagues, and at some point I felt there does not need to be a person doing this.

Deniz Kavi: You can just have a software solution from there. So basically trying to lower the accessibility gap, but also handling the infrastructure, the, software aspect of how you deploy computational tools like molecular design models

Grant Belgard: What are you most excited to build or learn about right now?

Deniz Kavi: Yeah, I think today we spending a lot of time on developer tooling to some extent. I think we have found that sort of– we work mostly with large pharma and biotech companies, and we found that these organizations have a lot of interest in how they do– how their R&D teams deploy AI to their use cases. And, ultimately these have served that demographic and like we are building a lot of tooling for that. So I think most exciting– most of what’s most exciting to me right now is a organization-wide, whatever, one to ten thousand person deployment of molecular AI tools for, serious drug discovery applications, things that will go to clinical trials or become tool proteins, what have you. Yeah, I think serving data scientists and AI folks to then deploy that to the whole organization is something I’m very excited about today.

Grant Belgard: What problem do you most want to solve for scientists?

Deniz Kavi: I think the general genesis of the company has this theme of accessibility. Historically, if you look at how software has worked in our industry, do a– you do a computational chemistry PhD, you’d take a four-month course on this one software product, and you become an expert on it, and then you are the person tasked with running all these workflows for your colleagues. I think we think there’s a spectacular amount of inefficiency in sort of the handoff in information or in convincing your computational colleague to do this work for you, like all the lobbying internal work. I think our general case is that of the increased capabilities of AI recently, it’ll be available for every scientist and not just specialists in, 20, 30-person data science teams, but the entire org will be the consumer answer of people applying the AI tools in our day-to-day.

Deniz Kavi: So I think very much the, having every scientist be empowered by computational tools as opposed to a realm of specialists doing it for them

Grant Belgard: What’s the backstory on the name?

Deniz Kavi: Yeah, unfortunately not a very exciting story, but we basically wanted it to be named after a living thing. We went through a list of trees, found tamarind as a, memorable fruit and a notable name, and then went from there

Grant Belgard: Where do you think the real bottleneck sits today? Is it model capabilities, infrastructure, workflow design, adoption, something else?

Deniz Kavi: Yeah, I guess I would be biased to say infrastructure. I think model capability is a credible thing to look at as well. I think the better the individual tools are, the more– the less certain-uncertainty you have, the more reliability you have on the individual models, the sort of more valuable the infrastructure becomes.

Deniz Kavi: Because if you have to tweak a bunch of knobs and find all the errors and evaluate against them, the model itself may not be very useful if you don’t know those exact details. But in our case, we’re seeing the molecular design tools especially become good enough in the space that, you can just point at a problem and apply it there and go from there, basically. Our general view would be having it be available to a scientist. I’m repeating myself over and over. So generally, very much the sort of focus of the company is how do you get somebody with an interesting idea for a target or a, drug program or a way to optimize a lead protein or lead sequence and how do you get them to convert that idea into, a molecule on the computer, basically.

Deniz Kavi: I think broadly the issue is not so much the hypothesis generation or having targets to go after. It’s more so like how do you– Once you have a problem to solve, how do you apply tooling to solve that problem? It’s like how we think about the, one of the gnarly-est problems in the space.

Grant Belgard: What does great software for life scientists look like for you?

Deniz Kavi: Yeah, I think we, from the start, focused on having an opinionated way to apply these tools in the first place. You– if you take a look at the papers of these there are a lot of focus on, like, how the model is trained, how their– how the inputs should look. Our view has always been basically, we built a sort of opinionated, structured way to consume those tools in a serious, drug discovery use case. So for structure prediction, complexes and docking is a big use case. For protein design a lot of papers come out and say just, we are creating a way to create new structures, and these are arbitrary, not really– these are not becoming drugs, these are not becoming antibodies or vaccines or what have you. And we build templates and recipes to make that work more effectively. That’s one avenue, is like you need to codify and build the best practices into your product.

Deniz Kavi: The other side is, I think, integration. So we have been very like the entire product from day one has been usable programmatically.

Deniz Kavi: So our API or MCP or agent offering is basically a way for any arbitrary code to call our product. So I think in a way, we can just go away in the background, just be done as infrastructure, and that just gives you flexibility to consume the tooling in wherever you want it. So it might be in your ELN, your LIMS system in your internal custom applications or your AI agents within Claude, what have you.

Deniz Kavi: I think those two are the main points of you want to build a opinionated, codified way, like a very practical-focused applications of the AI tools. And the other side is the how do you get it to– how do you get a scientist to use you where they want you to use them? That’s the two pillars, I would say.

Grant Belgard: What part of the scientist experience do people outside the field most often miss?

Deniz Kavi: I think amount of, uncertainty in everything you do, basically. Like I think some people think about the process as, I have an idea, I will do an experiment to fail or not fail that idea, and then that’ll be the end of it. And if it works, it’s gonna be a drug and it’s going into clinical trials.

Deniz Kavi: If it doesn’t work, I like toss it. But oftentimes, there’s uncertainty in the actual experiments you do and the computational tools you use. ‘Cause anything you do has some error bars associated with, and you have to believe and persist every time something is like going wrong where it might be a sign of it being interesting. Very rarely something in science is, immediately a useful result or like you, you even get data that like proves it in either direction. So I think that’s a very much something people not think about is you’re not really sure where you’re going when you’re in the middle of a process or like hypothesis in general.

Grant Belgard: What have conversations with users taught you that theory never would have?

Deniz Kavi: I think we don’t talk much about– People talk a lot about like where AI is interesting, where the applications are, but applications like are pretty broadly on what people think about as things that computers can do. I think if you think about the grand idea of a virtual cell, the sort of like a philosophical leaning of a research work there would be you want to create a thing that simulates the entire behavior of a cell in the computer. But realistically, that is like too big of a problem to be applied immediately today. So in some ways, I don’t have– I have a computer science background. I have a sort of, basic science background in, in relevance to that as well, but I’d never worked in a sort of specific R&D org before in a drug discovery context.

Deniz Kavi: And I have learned a lot about what the day-to-day problems are in a specific, in the context of a research program for a drug program what the use cases there are. So that’s things around what assays are most interesting, what sort of you can trust on, what you can trust what the sort of clinical or risks are before going into the later stages.

Deniz Kavi: I think that type of day-to-day experience is very hard to see. Certainly in academia, even if you’re a very strong scientist, you might not know the sort of how the applications go. And there is a meaningful difference in academic science and, commercial drug discovery-focused science, I would say. And, having not worked in that space, talking to– the privilege of talking to like pretty senior folks who manage these R&D programs and learning how they think about those use cases is very valuable for me.

Grant Belgard: How do you decide what to hide behind the interface and what scientists should still control directly?

Deniz Kavi: That’s always a changing structure for us for the most part. I would say people tell us that this point is a big value add for the platform, like maybe whatever tens of thousands of scientists, people will email us when they want a model added. I think we’ve always had a question of complexity versus abstraction, where, we have, say a few hundred person org in a company using us and also like some 30, 50 person data science team using us. And the interests of those groups are not necessarily the same. The scientist wants it to be easy to use. They can come in, use it, and then forget about it, but the data scientist might be, I want to check every single possible option. Generally, we’ve built our product in two distinct sort of ways.

Deniz Kavi: So the API, for example, is you can– you do anything you want programmatically, can write code to build on top of our product. And then nowadays, the API actually serves as a way to like abstract even further. So we found AI agents as a good way to guide the user to, you know… Most of our value from AI agents, like not even necessarily the reasoning, but just the hiding the complexity of how you consume a tool.

Deniz Kavi: That’s the one avenue. And the web interface is also there for the most complex way to Basically, anything you need for a given tool, you can use with another web interface. So we bifurcated our products in that way for different cases. And then generally just being close to the customer, and people tell us when they’re unhappy and, it’s a boring answer, but just people– you be very close to, your check-ins and regular meetings, what have you, where people will tell you what the, what problems they have and how– whether or not we’re able to solve them.

Deniz Kavi: Or if we can solve them, are they able to access that problem or access that solution to that problem are the main ways.

Grant Belgard: What does a trustworthy AI-assisted workflow look like to you?

Deniz Kavi: Yeah, I think generally the– I guess these are two different answers for, what we live in, the molecular domain, and also the sort of general knowledge work reasoning LLM tasks. Would say the molecular side, you just want wet lab data. I think we’ve seen a big explosion of the recent models doing very well in the sort of classical, one binder design or structure prediction problems, and much of that was launched by you can do experiment to see that this model works and like you can see the hit rates, what have you. I think that both added a lot of trust, but also despite not substantial changes in architecture, showed just that the models can be good and you can demonstrate that they’re good in that way. I think there’s still some questions there, like how much you can see from experiments. There’s– obviously, you can’t run a clinical trial to show that your model works every time.

Deniz Kavi: The molecular design side, I would say it’s the, just want to run experiment that can be as close to the use case as possible. Then surprisingly enough, like in some, most of the best performing models, they don’t really tend to correlate with a direct experiment. Like most of them are not simulations of an assay you would do.

Deniz Kavi: Many of them are, but not probably the most successful aren’t. That’s one. For LLMs, I think the

Deniz Kavi: main

Deniz Kavi: question is around, You know, to some extent, the tools LLMs have access to are not particularly strong. I think if you give the tools a bioinformatician uses to an LLM, there’s still some uncertainty and problems around that because of the sort of inaccuracy of the specific tools you’re using or the value of the, the reliability and the accuracy of the tools it’s using in the first place. If you give an LLM a protein design model, you now have two things to worry about, which is a faulty tool using another faulty tool, and you have to make sure that you have to find out which one’s the wrong part there. View for there– for that has been basically being very transparent and clear about what the tools it’s using are looking like and evaluating those results.

Deniz Kavi: Our LLM agents basically call our same– Like, whatever humans would use, the LLM uses that tool, and you can interpret it in the same way. So all the sort of inter– the abstractions we built help there. other questions are how good is the LLM in scientific parsing and understanding what’s going on. at this moment in time, the, Claude Op– Claude 4.6 and as of recording is 5.4, Thinking for GPT are reasonably good at not hallucinating anymore. I think questions tend to be where do you take the result of a paper at face– like what the result you’re citing at face value versus is it applicable?

Deniz Kavi: So let’s say if a model setting use a tool, the only tool it has access to for, let’s say protein solubility prediction, it just says protein solubility predictor, and it does not have any error bars or caveats or any details there. The model has no choice but to just, “I have this. This is the one thing I have.

Deniz Kavi: Let me just call this and give it to my user.” But in fact, maybe refusing to do something or showing its limits is probably a most more valuable thing there. Somewhat long-winded answer, but I should say, and showing uncertainty and then maybe even rejecting tasks when it’s on not confident in its ability to do them are two major points, I would say. And so I think the RLHF efforts on the LLMs are doing this well for other– medical advice or what have you. I think these will apply for science because we’re not really– we’re not trying to adversarially convince the model to do something that’s bad because there’s no benefit to having it do a ineffective in silico experiment, I should say.

Grant Belgard: How do you think about simplicity without flattening the science?

Deniz Kavi: Yeah, definitely a difficult problem. I would say the honest answer is like education and support is the honest or best way to go about this. I think we, we try not to limit capabilities of the tools we serve. To some extent we do, honestly, like for the, again, the, say small molecule design tools, for example. Like so many custom options you could be doing that it becomes untenable to use them for anything in the first place. So we try to build recipes of best, best– most commonly used use cases. I think the, again, the sort of open-ended AI models, like things like, protein design or small molecule discovery or even the LLMs themselves, they have a lot of use cases that can be disc– discretized into buckets of, this is my one use case.

Deniz Kavi: I want to do, docking prediction for small molecules, and I have a big world model that does it for me. I think narrowing those down to different tasks is the current approach people take. I think this is like unfortunate because if you can have a single model do everything, it’s quite, quite a more interesting experience and more interesting results most likely. But for the most part, it’s carving out chunks of the capabilities and giving them as recipes you can use is one, one approach. So you don’t really use– lose the capability there. You just become a bit more constrained into a way of doing things, which can be good if the model is not that good, is the– not that performant is the other detail there.

Deniz Kavi: For the most part, I think focusing on the where the models are good and like you can know that by being close to your customers, but also being close to the results in the first place with a sure review and then carving out those places to be individual tasks you’re looking at are the things we try to do.

Grant Belgard: What things are easiest to overbuild in this space?

Deniz Kavi: I think you can make a lot of analogies to like horizontal software players. So I think in some ways we resemble something like a Databricks or Snowflake quite a bit. We do model inference, training onboarding our custom tools, app builders most of those hold, but those products are also used a lot in the life science industry, and they can’t quite apply to them directly. the over-index on verticalized solutions for the industry. You think of yourself as a very bespoke product. There’s lawyers who do life sciences, there’s, procurement teams who do life sciences. And in fact, some of that might just be if you just take the horizontal version, like up and modify and optimize, that might be a decent way to approach it in the first place. You don’t need to reinvent everything.

Deniz Kavi: You can probably use the, the workflow builder, the pipeline builder for generalized use cases probably works well for the bioinformatics use cases as well. I think we don’t need to overindulge in our specialness and difference from the software industry or whatever the AI applications are.

Deniz Kavi: I think that’s the one thing I see a lot of is, you want to reinvent the tool hosting and say the bio– bioinformatics is basically, computer science, and then you don’t– a lot of the tools there, you can apply to the general use cases. So I think we– both from a people who gets into the industry perspective, like we hire a lot of non-biotech folks on the software side and then also what tools you use to build products are the– they don’t all need to be from scratch.

Deniz Kavi: They can be building on the horizontal solutions.

Grant Belgard: When new tools or models appear, what makes you pay attention?

Deniz Kavi: Yeah, I think the– surprisingly enough, I think the, to some extent, like the social media hive mind is like a decent predictor of like how much interest there is in something. That doesn’t necessarily mean it’s like the most accurate or the most interesting. Like some of our models are probably very, like very well known.

Deniz Kavi: People outside the industry know them, what have you, but they’re like basically not used. I think some level of interest tends to be a sign of in- of value. I think if something just goes out to bioRxiv and nobody talks about it, like it doesn’t tend to be the most valuable. I think that like in that case, like the most obvious thing is immediately valuable. The other side is these days, a lot of validation you can discretize and show for, especially in our use cases, in random molecular design use cases. I think if you show clear objective results there I think we’re doing a decent job of putting out benchmarks. There’s more to be done there for sure as well, like for other use cases. In my mind, it’s basically two, two avenues. One is, it can do something that was never doable before. I can make a fully– a protein completely from scratch and put it out and express in the real world.

Deniz Kavi: That’s meaningful scientific results. The other is it does something that was s- theoretically doable, but it was so much better than that, say 100 times better than that, that it can reliably get there.

Deniz Kavi: So again, going back to the protein design example, we’ve had a renaissance of binder design tools recently, and those are mostly because they’ve had, 100 times the hit rate of the previous generation of tools, and now people can actually use them and, instead of, they don’t need to have a huge V display set up where they spend a million bucks on the assay.

Deniz Kavi: They can just spend, whatever, $10,000 and show the results there. So I think it’s either increased accuracy and then the other side is just a net new use cases like in-invented by this protocol.

Grant Belgard: What trade-offs do you face between frontier capability and everyday reliability?

Deniz Kavi: I think you don’t want to be overambitious in applying tools to a given use case. Theoretically, there’s a model that claims to do everything you need to be doing for a drug discovery program. If you read the Wikipedia or you said the GitHub repo the definitions, and you just did them, you say you gave them to an LLM agent and then told them, “Don’t question these.

Deniz Kavi: These can do exactly what they claim to be doing.” And then probably like it will fail in the first step, but it’ll fail in the second step and the third step and fourth, con– successively. I think there’s a– incentive is always to publish like a decent result, and you just claim a… Maybe the most humble people say towards this task or towards, towards even a Jesse prediction on the computer. You do need to build the muscle of identifying what is there to publish a paper or what is there to show an actually or what shows an interesting result. I think that’s one question we have. Something that might be at the frontier might not be good enough for it to be actually practically useful.

Deniz Kavi: GPT-2 was on the frontier for a while, and it was basically useless for a lot of tasks. And similar things apply for our world, which is basically the incentive is to provide the most audacious result you can. But even– That might be true in relative terms. It might just be, it might be 100 times better than the previous stage, but that might not be enough from– If it never worked and it’s zero is zero.

Deniz Kavi: So just making sure that the individual applications are– I think restraints are, I should say in that note, restraints are a big thing we think a lot about where can a model be applied, what the actual sort of error bars around the results are, how can we codify that into a product where you can understand what the limitations might be. I think it’s just that there’s a lot of things that claim to be the best, and then the best is not good enough for a lot of use cases too.

Grant Belgard: What do outsiders misunderstand about making advanced computational biology tools genuinely usable?

Deniz Kavi: I think the view or the understanding is basically that everybody’s aware of what the protocols they want to use are and like what the best, what the hit rates are, what the– People think about, computational modeling tools as the same as they would think about like assays, where there’s some clearly defined set of worlds.

Deniz Kavi: You can buy a recipe or have a recipe built internally and just apply that, and like it’s, 90% of the time it’s gonna be the same thing, except some modifications you have to make, what have you. I think the– both the world is changing so quickly that like very few scientists are actually on top of what’s the frontier.

Deniz Kavi: I think like we do that role informally by just giving guidance on that and providing educational materials there. For the most part, yeah, the limitations are just, people don’t really know what works as a community, basically. So like I think it’s very much in flux, and that might maybe be stabilized like every couple of weeks and then we go off the rails again. The question basically is: scientists know what they want to do? Do scientists know what works well? And, the criticism of the science community is a sort of general observation that the world is moving too quickly and the incentives are, or incentives are such that everybody claims state-of-the-art for everything, and it’s hard to be on top of what the direct applications are.

Grant Belgard: What first pulled you toward the overlap of software and biology?

Deniz Kavi: Yeah, I actually– so coming into Stanford, I had done some work for doing natural language processing work. I very wrongly predicted that sort of transformers would not be a good application for life sciences. And so I was doing some older text translation and summarization, sentiment analysis type work, and I just assumed, “Oh, these are not gonna be useful tasks,” and, I’m gonna look at another application. And I got into biology from trying to find cool applications of AI tools as I was going into a pre-science degree. And in reaction to that, I got a job in a drug discovery context at a lab at Stanford. Yeah, I think it was very much I was trying to find an interesting application for computer science, which sort of led me to biology.

Deniz Kavi: I’ve heard this from a lot of mathematicians as well, where the modeling of a complex system tends to be an interesting problem as a mathematician. You become a bio-focused mathematician or what have you.

Grant Belgard: Was there an early moment when you saw a gap between what scientists needed and what the tooling actually allowed?

Deniz Kavi: Yeah, I think the most obvious for me personally, the earliest sign was basically that I was hired for a full-time or full-time job doing this for my lab mates. Somebody– I think to this day we see a lot of this as well, where even large pharma– A director of data science at a big pharma company who has 15 years of experience, half their job is just somebody sends them an email and says, “Can you run this in silico workflow for me?”

Deniz Kavi: And then they do it for them, but it like takes a few weeks, what have you. I was basically doing that for my job. So we had a protocol for finding targets for peptide candidates in a, in silico way, and that was interesting for applications of my lab, and that sort of led to some other separate work that was a cell systems publication. And that was out there in the world for people to use, and then nobody in the world could use it, basically. They would have to have me do it for them. And that was the genesis of the company in general, was just solving that problem of, I have an idea to apply this, but I have no idea how to actually, use the tool to apply them to a real use case. And that’s how the company came to be too.

Grant Belgard: What did being close to research environments teach you about how science really gets done?

Deniz Kavi: I think that there’s just, the most part, being close to scientists makes you appreciate both how valuable their work is, like how much how hard they’re working for these use cases. But the other side is, I guess I’ll– going back to my previous point, like a lot of uncertainty around how well do you think your idea will work. The experiment you do might fail, the computational tool you use might fail. There, there might be some, side effect you’ve never known about and, that sort of– you have to embrace that as a researcher. And I think researchers also think about the world in this way. I think it’s more, more importantly than typical day-to-day experience is a very strong skepticism of everything that comes their way.

Deniz Kavi: Like anything that could be interesting or useful, people are, just by training of the job, I think very much skeptical and often negative on these things, which I think is an interesting habit to have, and I think it has its pros and cons. I think the temperamental approach is like what the temperament of a scientist is, or life sciences, I should say, is for what the future holds, quite interesting, where they simultaneously have a job that is much, I’m in the frontier.

Deniz Kavi: I’m trying to invent the next thing, but also I am also very skeptical and generally negative on everything that might be interesting. And you have to fuse these ways to do the process of science basically which is– I thought was quite interesting.

Grant Belgard: What made the founder path feel right for you?

Deniz Kavi: Yeah, I think in our case, we built the first version of Tamarind for my lab as like a side project with my co-founder basically, and gave it to my colleagues. And then I posted on some random forum about, we have this use case, these alpha fold on the web, and web, and that led to several hundred users.

Deniz Kavi: So there was a lot of demand for academic users to– who like felt this in– problem of how they do computational tooling. that sort of felt strong enough as a pool of interest from users that we felt we, we should do this full-time and not spend our times doing this as a side project in school. And I think that was the primary reason. Like for the most part, in my mind, always not really thought about being– Like I always wanted to be a scientist or a researcher, and the approach in my mind was: What is most valuable to the world right now of the things I could be doing? What is the fastest path I could get there? And I also was very much pulled towards this idea of, I can move faster in a quick, quicker way than typical timelines you would take.

Deniz Kavi: Instead of doing a– four-year undergrad and a PhD and another, postdoc and many years of working at a biotech company, I now get to basically do the job of a executive at a relatively serious, bio-software company and just just by having the virtue of having built this thing as opposed to getting approval from everybody else on top of me is the sort of primary motivation.

Deniz Kavi: So combination. We had a thing that was working that we wanted to make it more serious. Just the waiting and the having to play the games of getting, accredited and approved and going better and better over time by some other institution was not very appealing to me. And the other side was basically for us to realize this idea of how we get this to– these tools to be used by scientists more effectively and at more practical use cases. It do- it has to be a company as opposed to a nonprofit or academic research project because the incentives are not strong enough for those use cases or for those methods of applying this problem that we had to make it a company, basically.

Grant Belgard: How did you learn to tell the difference between an interesting technical problem and an important user problem?

Deniz Kavi: Yeah, I think I was somewhat commercially minded which I didn’t realize at the time, but I’m basically a salesperson now, so I’m more aware of it today. I think many interesting technical problems can be commercial or user problems as well. The question is, we’ve been asking for people to pay us from day one, basically, as the one side. I think if you keep it as a free tool that everybody can use, it’s a– have too many nice-to-have type problems. I think you want a investment from the user of their time or their money or their feedback, what have you, where the problem has to be severe enough in their day that they would take some risk, basically, is how we think about that. I think I have a decent way to emulate that in my head today. I can understand what the workflow looks like from just talking to somebody about their day-to-day.

Deniz Kavi: But for the most part, do inference scaling for AI models. The condition for that to be useful for users is that the AI models are useful for them, and we had to validate that for that use case. my finding was, if you are willing to put something of yourself on the table, like risk something of yourself or time or money, what have you, is the main beneficiary of user solutions. And then technical products are often, you know– I think engineers are very much drawn to this case of “I think this is interesting. It’s a thing to solve,” but there’s no real thought about what it means to solve a problem as opposed to research curiosity. Yeah, that’s my general approach is getting commitment from the user is the main thing to look at.

Grant Belgard: What surprised you most about turning research-driven insight into a company?

Deniz Kavi: I think I was broadly surprised by how disconnected academia is to some extent from industry. Like theoretically, all the IP from these companies, all the biotechs comes from industry. It gets spun out of labs, it’s the professor’s research, what have you. But in many ways also, there are dedicated industry research conferences.

Deniz Kavi: The problems you tackle are quite different. Yeah, I think a lot of the basic science is sort of– you know, obviously, there’s some actual practical R&D work also happening in, in academic settings. I think very much the context was, biotech and pharma think about the world in a, series of drug programs or like potential assets you can acquire versus, research will be more platform style.

Deniz Kavi: You want to get a– I want to make the machine that makes drugs as opposed to I just want to make a drug and get– create a company around that or, provide a way to create that drug in the first place. Generally, I would say industry is much more open to investing resources to make things more convenient and efficient versus, academia or research in general can be, I will have an undergrad do this for five, 10 years and just keep going through and churning through them.

Deniz Kavi: I think there’s a lot of benefits to some problems that can just be like thoughtfully thought of and then ground through for many years or many long amounts of time. But it also means the timeline for everything you do in academia is very much shortened by what you can do. Like academia or research in general can be a source of commercialization. But if you approach the world from a perspective like I am doing something that’s very hard, it’s impossible to do, and I’m just like seeing how close to impossible I can get, that sort of somewhat grounds you in your ability to do interesting sort of groundbreaking work in some way versus if you think the thing you’re doing is very hard, but it’s like possible and like you have a timeline for that.

Deniz Kavi: I think, projects that are set up in a way where it’s gonna be I’m gonna do this in the next two years versus I’m gonna do this however long it takes tends to be an interesting dichotomy there.

Grant Belgard: Where did you underestimate the human side of the job?

Deniz Kavi: I think for the most part for us it was adoption. So there’s, one– even once you like c-close a contract or a customer signs up, like you use it within their organization there needs to be a deployment process. You need to be able– can you do basically lobby and convince functional heads, get their teams to use us? More than I… And think we have a more rational sales process than we most other industries do. Like we don’t do these steak dinners or whatever. Like it’s mostly a case of proving product value and then scientific interest and value there, and getting people to trust you and be the source of expertise, I think pays a lot of dividends.

Deniz Kavi: And that is to some extent bounded by knowledge, but also the cases, you want to be writing a lot, talking to people a lot, and that sort of tends to be the way you become a brand or a known entity in both attracting talent or customers or what have you, is very much driven by these human social problems. And then I would say in the deployment side, it’s very much driven by like political problems, like how you get a team to adopt you and how do you get a larger team to adopt you and how do you get them to use more or pay more and what have you are… ideally, you’re convincing that they’re– you’re spending more on Tamarind is a sign of getting better science, but also more adoption means more people have access to these tools that are more practically useful for their day-to-day. I think I’d never really done any of these tasks.

Deniz Kavi: I think I’ve become a salesperson, as I mentioned, and learning that on the job was very exciting.

Grant Belgard: Looking back, what moments feel like the real turning points?

Deniz Kavi: I think we– So the company’s about two and a half years old today, or two years and a few months. And the first year, we basically didn’t really know what we wanted to build. We had this idea of a, web interface for all the models, and you can use them easily. But the models weren’t good enough yet.

Deniz Kavi: We weren’t sure about what the applications would look like. What the– is the use case, production agents? Is it experimentation? Is it like proteins or small molecules or, how much of the scale part matter? And we had to invent that with our customers, and that was a pretty meaningful thing for us to go through over time. In that sense to invent a category basically. It was the most surprising experience for us was to figure that out. And unlike, if you’re starting a CRM company or like a pretty typical software company, you can just check what’s happening in the world and you do it– you could do it again, but better. In our case, it was very much a case of how does this product happen ever and how does it become actually applied? I think for the most part, there’s been a lot of efforts in this direction, but none of them have gotten that much actual usage.

Deniz Kavi: So we had to invent what the of that product would be.

Grant Belgard: What skills compound fastest at the biology software interface?

Deniz Kavi: I think I would put just like being an information sponge as a very valuable skill to have. I think people underappreciate how much– if you just know a lot about the use case you’re interested in, if you know a lot about the customer you get to or just what the product you’re building is. Oftentimes, again, the world is changing very quickly. You want to be on top of everything. My view has always been I’m gonna spend all the time I have in the world passively learning what I can and being as close, curious, and interested in people as much as I can be. This can be, whatever, social media posts, YouTube lectures, or, it could just be a scientist you want to talk to. think a benefit of being in the world is that you get to talk to experts all the time.

Deniz Kavi: And I think if you’re doing a computational science PhD, and all you do is like you re- you write papers and you just publish them and nothing happens afterwards. Talking– like having a person that uses and is depending on you is a very strong incentive to learn in general. And the ability to learn is a very powerful thing to have. And putting yourself in a structural position where you can learn a lot is just quite valuable. The PhD obviously like you’ll learn a lot there, but the things you learn are very much how do I invest– advance my general PhD research theme versus, people are relying on me that are quite important in their organizations.

Deniz Kavi: What can I do to just solve their problems or go another avenue on the learning side?

Grant Belgard: What’s worth learning deeply even though the tools keep changing?

Deniz Kavi: I think honestly, the– one of the strongest– one of the biggest gaps we have is like very strong core AI folks. I think you can see the incentives here, like if you’re a really good AI person, why would you want to work in biology? Basically, like you can do the core foundation model research, you can do some interesting robotics work or do computer vision, what have you.

Deniz Kavi: You’ll probably be paid more. There’s a couple of companies I think like bucking this trend. There’s a few like good research organizations, but I think the core foundational computational skills are quite important to learn right now. And there’s a lot of these getting replaced by the old LLM coding agents, but I think having that technical fluency is quite valuable in that, I think many bioinformaticians have become basically people in the interface of software and techno– software and biotech, is they become masters of none. And that is not a great place to be for most.

Deniz Kavi: You’re being bought– you’re being hired for a specific skill, or you’re starting a company for a specific skill, and if you’re okay at everything you might be a decent cog in a machine, and you might just be helpful in an organization, but you’re probably not gonna move the frontier of that field forward very much beyond that. I think founders become a lot of generalists. I think you do need to be a generalist. I fundraise and sell, and also you write papers about like our benchmarking results, what have you. And to some extent, being an expert in something that is to computer science and then applying that to life science problems can be a very valuable way to approach this.

Grant Belgard: How should early career people think about breadth versus specialization?

Deniz Kavi: I think my general concern about breadth, especially in our field, is that you are putting together a lot of things where I’ve taken some classes at Stanford that were basically surveys of the field because you like want to simultaneously appeal to computer science folks. You want to learn about bioinformatics.

Deniz Kavi: to apply t- to apply the computational scientists, learn about more computer science. And those courses are like useful, but they also become basically surveys or like very high-level walkthroughs of the field. And then those don’t tend to create a thoughtful scientist in and of themselves. I think my view in the world in general, like not even beyond this field, is if you want to be an expert in something, and then if you are an expert in one thing, you can then become an expert in another thing and go from there, as opposed to trying to do those things simultaneously. To some extent, I have some concerns about the… I have skepticism of inter-interdisciplinary degrees, for example, in that direction where like classes or majors that prefer to be doing these two, two things simultaneously.

Deniz Kavi: Because if you’re trying to learn two things at a time and then combine them together, the degree program like makes a theme or makes up a way to connect those together, but not necessarily a way to actually be a good credentialed expert for that use case. My view would basically be, become the smartest person you can be for one field. I would probably start with computer science, honestly, and then see where you can apply the science from there, is my general view.

Grant Belgard: How can people get close to real scientific pain points instead of building in the abstract?

Deniz Kavi: I think honestly, talk to people or put yourself in a place where you have to do it. I think in some ways one of the best ways to learn about the economics of a therapeutics company might be to start a therapeutics company. I don’t know that I would recommend doing that for everybody just to learn about that stuff. But I think I really believe in the power of being forced by your environment to do something hard and learning those things on the fly. I think if you’re reading papers about how the drug discovery, this process is done, like it’s just hard to be in that mindset. Like it’s even hard to motivate yourself to read those things, right?

Deniz Kavi: ‘Cause like it’s not gonna be immediately useful to you. So I think put yourself in an uncomfortable situation. I guess in the like general case, I have friends who have applied for jobs in places that they’re, they were definitely not qualified to do, and they offered geo service for free for them, and if somebody responded, they would, they have to make that work in the next one or two weeks, and they respond in that way. I think putting yourself in an uncomfortable spot like forces you to learn something, and like that might be a bit embarrassing also if you fail. this is how I think about those things.

Grant Belgard: How should scientists decide among various career paths?

Deniz Kavi: I think it’s just, it’s a question of, in my mind, I optimize on what I think is gonna be most impactful. The most people or has the most meaningful scientific sort of core that grows into something else, is how I think about that. Part of the reason I became a founder was very much in this context of if I did a research route, might be a good scientist like five or 10 years from now, but if I’m the median scientist, I’m not gonna be very important to the whole of science. And I’ll have a lot of papers with have 20, 20 citations, don’t really go that far was what I was concerned about personally. You can be a great scientist. I think academic research or, re- general research focus is very important, and some people will do very powerful, impactful jobs there work, work there, I should say. My view has always been what are you optimizing life? In my mind, that was impact.

Deniz Kavi: So however many people I can impact, I would like to optimize for that, was how I thought about that. So if your belief is that, the s-science in– of the next generation of, say virtual cells will be in academia, you should go all into that and then become a PhD and a professor and doing virtual cell work. If you think it’s gonna be applying some AI tools to therapeutics work, you can start a biotech company or start your own company to some extent. Or the other side might be, I think the most important bottleneck is clinical trials, even though I have this life science background, let me go try to find out what the clinical trials can be solved with. Yeah, I think so find the thing to optimize that you think is most valuable and go for that.

Grant Belgard: What mistakes do smart, ambitious people make when they enter this space?

Deniz Kavi: I think the… I think the mistakes smart people make are always almost the same in independent of the field. I think oftentimes the questions are you try to learn from somebody else, become a clone of what you think is a good thing to be doing. I think it’s a– for the most part, it’s like a decent way to live your life is to copy someone who is better than you and pretty close proximity to you. I think that’s reasonable, but to some extent, you do want to take a leap at some point. You do want to take a risk as much as possible, and that doesn’t necessarily have to be like, dropping out of school and starting a company, but it can also be something like, going beyond replicating somebody else’s work, I should say.

Deniz Kavi: I think we set ourselves like heroes and people to admire and work to replicate and to be similar to, but that is– that gets you only so far in what you can accomplish. It’s helpful to have knowledge of history and what things happened before, but you probably don’t want to be living in the shadow of somebody else or someone else’s work as much as you would in an empirical context.

Grant Belgard: What habits help you stay grounded when the hype cycle gets loud?

Deniz Kavi: I think I just might say talking to customers again. I think sometimes I’ll– people will tell me their negative results, so I think it’s quite go– There’s a lot of ex accounts who just talk about how biology will be solved next week. And like I am quite like– I think I’m very much exposed to that technology world when I’m quite excited by the technology in the first place. But once you know what’s happening in a field, it becomes much less exciting for the most part. I think, I suspect, people who do the core AI research for LLMs are probably much more conservative than the people making products on top of them. They’re probably both very excited and ambitious right now. I think we’re actually yeah, I think ke-keeping yourself in a place where, you know, people whose job it is not to sell AI for their company, but people whose job it is to apply it to their use case.

Deniz Kavi: This is very AI-specific, but also can go through most everything else is, when you open news-newspaper, you see the headlines about science, and you think they’re all bad, and you read all the, headlines about politics, and you think they’re all credible.

Deniz Kavi: I think it’s one of these things where you just want to be close to the source as much as possible if you care about the results from a given category.

Grant Belgard: What question do you wish more people asked you?

Deniz Kavi: I think a lot about why there have not been great software companies in the life sciences over the, say past 50, 20, 30 years. I guess to answer that, I think the value of software was somewhat limited. most of R&D spend in biopharma goes to development, like it’s, assays, experiments, headcount and then like a bunch more of that money goes to clinical trials and all the commercialization efforts for like obvious reasons, because it’s much more expensive. And I think a lot about like why we are different in this world where, basically our belief would be AI is sufficiently new that– or powerful that compute will be a more meaningful spend on the biotech side, and this company can be a whatever billion dollar revenue company in the next five, 10 years. And, I get asked this from like VCs a lot, but not so much by products and practicioners a lot.

Deniz Kavi: And I am curious how folks think about the tools they use, like how viable and sustainable the companies themselves will be. Like in many ways, a lot of these companies get bought by private equity, they get merged into some other thing, like the product does not improve over five, 10 years. And, if you’re scientists who use these tools, don’t think too much about what happens. It feels like whatever happens to the company happens to them also, and like they don’t really get to control what happens. And I’d be curious about how scientists think about what software tools they use exist as sustainable, long-living companies versus, a thing they bought for a year that might disappear next year.

Grant Belgard: What belief do you hold more strongly now than you did a few years ago?

Deniz Kavi: I think I’m a believer that AI will be valuable drug discovery. I think like this is a– I don’t see the starting company as a more open-ended question, because of the bottlenecks in clinical trials. Again is AI the right place or the– is drug discovery the right problem to apply this even in the life sciences world where– and what if you’re optimizing clinical trials, you’re like shaving one year off a clinical trial? I’ve become very convinced that at the current rate of getting better, the models improving, will have, net new assets added to every biopharma company’s portfolio of assets that will be, driving revenue for them. I think like we think a lot about, science outcomes like approvals and clinical trials, but I think, AI will make more money for pharma companies is on the drug discovery side by making more drugs for you to get to clinical trials and then approvals eventually.

Deniz Kavi: And I think this will almost inevitably will happen in the next of years or very soon, basically. I think, the adoption is there, there’s interest, but also I think the results will be speaking for themselves very soon.

Grant Belgard: What about the inverse? What have you changed your mind about in the last few years?

Deniz Kavi: Yeah, I think the– I am generally actually averse to science for like, goodness sake. I think you do want to think more about what the outcomes will be. Yeah, I think revenue is a core metric when you’re trying to make medicine, think about the commercial results. But to some extent you are serving tooling for an organization, and I have not thought about, I think about science as an interesting, valuable way to impact humanity and be good for the world, what have you, and I believe that still. But what you do day-to-day on– at your job is basically the same, whatever job you do. Like some– or between the roles I can do anyway are very much the same. And you want to pick your battles more carefully what you apply yourself to. I think I used to think that, you want in a field that is good for the world necessarily has to be strictly a very clear value add.

Deniz Kavi: if I went back to my, yeah, to three years ago and I start– just starting a company, I might just tell him, think more clearly about do you want to work in science forever?

Deniz Kavi: Do you want to make a bioinformatics software company? To some extent, you don’t necessarily want to be making a plan for yourself without knowing the requirements for that plan. So in this case, it would be, let me rephrase what I was trying to say, which is basically is the sort of aligning core thing you’re motivating your acts by?

Deniz Kavi: I mentioned previously you want to pick a job or a thing– or one singular thing that you’re optimizing for, but be more thoughtful about what that thing is and not just which fields are is good or perceived to be good by from a fuzzy, open-ended sense. And I think there’s a lot of value to be added for clinical trials, like a lot more boring even like payroll or like better health insurance, what have you.

Deniz Kavi: That type of thing is still quite valuable, and I would not– I would have looked down on those companies beforehand, where I’m now much more agnostic to humanity. This is a singular sort of effort to make the species better, and whatever you dare can be quite valuable.

Grant Belgard: What boring looking skill turns out to matter far more than people think?

Deniz Kavi: I think generally remembering things is quite important. I think like making structures for yourself to remember people’s names and where they work and what their, use cases are. I think it pays a lot of dividends in getting to know people effectively. If you keep seeing the same person over and over again, you can talk to them more effectively.

Deniz Kavi: Yeah, I think taking a lot of time to remember other people and also the specifics of their work or like what they might be related to you is quite valuable. We’ve seen this a lot in hiring, where like keeping track of somebody for many months, like just eventually converting them to a part of the team or a customer in our case.

Deniz Kavi: Or it might just be scientific collaboration or working with a partner on a specific program. I think just being very– taking interest in people and being thoughtful about keeping that interest in your– like in a structured way. Maybe you should remember from the top of your head, maybe you should write it down for yourself. I think paying attention, is the general theme here, is like purposely and spending a lot of time paying attention is quite valuable.

Grant Belgard: What are you still trying to get better at?

Deniz Kavi: Yeah, I think I spent a lot of time living and breathing with scientists, and I still do most of the time. I think we are as a company looking a lot more like how we can get this enterprise level adoption question of, a mature company in our space should look like, nine-figure revenue per customer. And that means basically everybody at this company uses you and like for a wide variety of core tasks. I want to become a better– I run a what, 15, 20 person company today. I’m not like a huge organizational figure. They’re all in the same room, basically. My– I want to learn about how our customers think about systems, how they think about the organizations of their large organiz- companies, like making those func- machines function well.

Deniz Kavi: Having managed a small number of people so far is a very complex problem, and both from understanding of what product we build next, I would like to be understanding of how the organization behaves and reacts to those things, but also running my own team more effectively. I think so I have basically over the course of the company, I’ve picked up skills I’ve never had and I was quite bad at the start, and I’ve become pretty good at.

Deniz Kavi: I think the next thing I want to be best at is understanding the behaviors and intents of a large organization and building products first and foremost for those organizations, but also apply those learnings to my own

Grant Belgard: What would be exciting for you to see happen in this space over the next few years?

Deniz Kavi: I think if we got 50% of the way to what the models claim they’re able to do, honestly, it would be very exciting. I think that would be a pretty meaningful proliferation of these models for different use cases. The approaching design is reasonably good, but if we got to I guess whoever those sort of hype people say on Twitter about these models, if they become real, which I think is like within the bounds of the we get today, would be very exciting. I think ultimately Tamarind is a index on the belief that, models will be getting better for molecular design applications, and it’s obviously good for me that the models get better, but also for the value of how we can produce more medicines more quickly, more effectively, and then even before that, just producing new medicines, period, so you can have another list item in the list.

Deniz Kavi: So I guess beyond the more boring and the models get better, but if they got as good as what people, some people think it, they are at right now would be the main thing I’m looking forward to.

Grant Belgard: Last but not least what would you like our listeners to take away from this conversation?

Deniz Kavi: Yeah, I think the general thing is you don’t really need permission to do things. You can start a company, you can work on a research project. That doesn’t mean that it’ll be easy. People will be rejecting you very frequently. But being in an environment where you are constantly rejected, being very– in an environment where you you’re punching above your weight is a very valuable place to be. I think people go and follow existing tracks a lot in life, which is very comfortable. It’s not pleasant, but what you need, what you’re supposed to be doing. I think you can go above your station and if you can hold yourself long enough there, you probably will ascend to a larger point. I am a not very long career– I did this out of school, and I started a pretty meaningful company still, and, much more to be done for sure. It’s a pretty, pretty small company still.

Deniz Kavi: And I think that the way– The limited success I have had came from being– putting myself in a place where I did not feel welcome or I felt like I was not supposed to be in this place.

Deniz Kavi: And oftentimes I was not supposed to be. Like, I was clearly not qualified for some of the things I began doing. And then just being in that situation makes You rise to the occasion when you’re put in an occasion where you have to be responding to something very hard, and putting yourself purposely in a place where you have to do something quite hard or at the time not possible for you to be doing is quite valuable.

Grant Belgard: Dennis, this has been fun. Thank you for joining us.

Deniz Kavi: Yeah, thank you for having me.

Transcript of Episode 86: Zahra Jawad

Disclaimer: Transcripts are automated and may contain errors.

Grant Belgard: Welcome back to The Bioinformatics CRO Podcast. Today I’m joined by Dr. Zahra Jawad, founder and CEO of Creasallis, a biotech developing antibody engineering technology aimed at helping antibody-based drugs penetrate solid tumors. Zahra’s career has spanned antibody engineering and innovation roles across biotech and pharma, and today we’re talking about what she’s building now, how she got there, and what advice she’d offer to others building their own path in science and biotech. Zahra, welcome.

Zahra Jawad: Thanks for having me. It’s a pleasure.

Grant Belgard: So what are you building at Creasallis right now?

Zahra Jawad: At Creasallis we are really trying to improve how antibodies therapeutics penetrate solid tumors with our antibody engineering solution. It’s something that I came up with, during my sort of 20 year research in antibody engineering that I thought is really needed in the field.

Grant Belgard: What problem are you most determined to make meaningfully better?

Zahra Jawad: Antibodies are, have been amazing therapeutics and it’s really changed a lot of patients’ lives, but in reality, they’re so large in, in their molecular size that they don’t penetrate solid tumors very well at all. If you look at what a tumor is it’s got like very tightly packed cells with a very abnormal vasculature around it. And the antibodies, once they leave the blood vessel, really don’t diffuse very much into the solid tumor. And there’s been very few solutions in trying to improve that. One of them is just making the antibody a lot smaller, but when you make it a lot smaller, you have another problem is that it diffuse, it just filters out of the kidney very rapidly, and so you lose this half life and there’s always this dichotomy between half life and versus tumor penetration that no one’s really solved. So what we’ve done. Is really taking inspiration from the space industry.

Zahra Jawad: And like a rocket takes off, like we saw recently into outer space. It’s got all this fuselage attached to it, but when it actually reaches the atmosphere, drops a lot of the fuselage that’s needed. We do the same thing. We re-engineer antibodies so that they break into small fragments and only the really critical parts get into the tumor cells rapidly. And so that’s. We get that sort of size change in the tumor that we’ve engineered in.

Grant Belgard: What are the biggest buckets of work on your plate at the moment?

Zahra Jawad: The biggest problem always in therapeutics is translating what we see in the lab and what we see in animals into real clinical, meaningful data for patients. That’s really what keeps me up at night more than anything. How can I make sure what we are doing in the lab will really benefit patients at the end of the day. And how can we capture that diversity in patients, in terms of their medical diversity, their ethnic diversity, gender diversity, genetic diversity, and still be able to give them a meaningful therapeutic that isn’t gonna be so personalized that it’s gonna cost phenomenally. So the economics versus the patient benefit.

Grant Belgard: Can you tell us about your journey as a founder?

Zahra Jawad: So yeah, I probably have a bit of an unusual journey because I was working in biotech and pharma since about 2006. After I finished PhD in some postdocs, all in protein engineering, I was really happy and content in my, my role. I was learning so much from the people around me and how to make therapeutics. And was very satisfied. And I just moved naturally up the, career ladder from sort of scientist to senior scientist to investigator to team leader and then a director level in charge of innovation. And I was really content in my job. I think the turning point for me was when somebody in my group developed a stage four lymphoma. This is when I was a director. I was in charge of innovation, antibody innovation, and I think, and she was young. She was 32 and she made a full recovery.

Zahra Jawad: She was a superstar, but she said I think she was on her fourth cycle that her mother wasn’t able to come with her to the hospital for that chemotherapy and it was actually antibody therapy that she was getting. I thought I’d just be a good boss and take the day off and go with her to the hospital. I guess that was really the turning point for me ’cause that was the first time I’ve ever really been in contact the real patient experience of people, patients taking antibody therapies real life, in an outpatient department where there were several, maybe half a dozen people there all in the same situation. And when I realized we had a problem with antibody therapeutics, that they’re actually really toxic for the patients. And I was and that really affected me, impacted me quite significantly.

Zahra Jawad: And that’s really where my thinking changed to how can we improve antibody therapeutics to make it safer and less toxic for patients? And that’s where sort of this idea of changing the size of the antibody within the tumor came about. It’s not easy for someone in industry to have a startup because I don’t have the tech transfer support from a university. I don’t have- any I, any IP actually I own is owned by, or my, entirely by my employer. So I couldn’t generate any data. I actually never spoke at all about the idea and I sat on it for about three years. It was then that I joined a startup called Bit.Bio where, at the time the CEO convinced me to come and join them to help them get projects through the door. And I’m like I’m a scientist. I’m not a project manager. It was like yeah but that’s why we really want you, ’cause you understand the science and you understand scientists.

Zahra Jawad: But I did say at the time I’m a scientist and I like developing therapeutics. And the CEO was like, that’s fine, we have ambitions for that and we’ll give you a scientific role within the company. So I decided to join them, not thinking at all about my idea of penetrating antibodies and really threw myself into that role, found myself after a year being actually, a little part of me died that I really wanted to be close to the science again. But when I wanted to get back into the science, the company really needed me in project management more. And so I was quite happy to do that, but I still missed the science immensely.

Zahra Jawad: And it was really with the support of that CEO, where him and many others who encouraged me to think about like my original idea of penetrating antibodies, very generous in letting me explore that while I was in employment and eventually got to a point where I generated enough interest to be able to raise a very small pre-seed round that enabled me to leave my job and do my first ever experiment around the idea. So I was a little bit unusual where I managed to raise a little bit of VC funding, really just on an idea and, with no IP and no team. But it was really the great support network I had around me that enabled me to do that.

Grant Belgard: How has the time you’ve spent at the bench changed since then?

Zahra Jawad: What in that year that I was in Bit Bio a couple of years?

Grant Belgard: Starting from then to now.

Zahra Jawad: Not much actually. I think like when the process that we drive for therapeutics has been similar in the last, I’d say 20 years of my career. people who develop therapeutics don’t like to change their process very much. And. I just think there’s so much opportunity to go faster, to streamline a lot of things with a lot of automation coming online, AI enabling. I think it’s wonderful. So I think like I would say the only thing that changes, I moved a lot faster than I’ve traditionally moved. Taking all the experience I have, knowing where I can take a shortcut and where I couldn’t. So it was really focused. I, as soon as I quit my job I went into the bench and I filed a prophetic patent and my lawyers warned me I have 12 months to get everything in there. And I’m like, to what point? They’re like to in vivo proof of concept you’ve gotta show in a mouth. That it works.

Zahra Jawad: And I’m like, okay, I’ve got 12 months ready, steady go. using a lot of the knowledge I knew of where, what really moved the needle to get data through. I just got that all in, in just over 11 months and was able to comfortably file that patent, with all the data needed.

Grant Belgard: Can you tell us how Creasallis is using AI?

Zahra Jawad: Oh, I have like several chat bots going at the same time. So we don’t, AI isn’t integral to our engineering. But it’s integral to what, so there’s several bots that I have not only to run the company, I really do believe we are very small with, we’re really four people, I do really believe in using bots to streamline and automate a lot of operational work processes that I don’t need to get involved in. So definitely a big fan of that. But I also have a research twin or a buddy of which I use to scan the literature, interrogate X and interrogate discussions on social media to challenge my thoughts, to come up with new ideas, new targets use it for recruitment to try and find the right people. Everything. So it is quite integrated, I would say within our, the way we work every day. AI is used every single day in our organization.

Grant Belgard: How would you estimate that’s increased your productivity? Pre AI, how large would your team have to be compared to how large it is now to achieve the same level of productivity?

Zahra Jawad: Yeah. So really the only roles we haven’t replaced with AI are the people at the bench Although a lot of what we can do in terms of analysis can be streamlined with AI even with just off just with chatGPT and Claude is probably more than enough for us to do all the analysis we need to. so I think. Typically in a biotech company, I would say we used to run antibody therapeutics programs with, I’d say a minimum of four to five FTEs at the bench. It’s not operations and we are doing two programs with three FTEs at the bench. So we are definitely able to get a lot more done with a far, maybe half even, maybe. A quarter of the number of people? Operations is completely AI. All the sort of mundane tasks of emails, financing projections, report financing reports, putting things into board decks social media. It is all automated really.

Zahra Jawad: I don’t really want to I want to be, think, I want to be using AI to converse, most important strategic decisions rather than spending my time filling in Excel tables.

Grant Belgard: Makes sense. So what, what would meaningful progress look like for Creasallis over the next few years?

Zahra Jawad: Oh, racing to get to the clinic as quickly and smartly as possible on a very lean team. Definitely our plans are to probably, we are probably going to be filing an IND or some sort of regulatory submission end of next year. So that’s the next 18 months. And in that, then after that, I would really wanna be able to see some sort of safety phase one data around our concept with some early efficacy readouts. Um, there. So we do have, we have several options. We have options of. Taking all these products all the way, like improving, making the next generation of antibody therapeutics that are better than the first generation and getting all the way to market, what really my vision is that every single antibody is going to need our technology. The fastest way of doing that is partnering as much as possible with pharma companies and. Uh, to do that.

Zahra Jawad: So that’s my vision is that Creasallis’s technology will have to be in every single antibody moving forward. And the fastest way to achieve that, in my view is partnerships. So that’s really what I’m trying to focus on more, is to make that, that next generation of antibodies through as many partnerships, not just with pharma, biotech, startup, research institute, CROs, everything, like exploring that all.

Grant Belgard: What makes a partnership worth pursuing?

Zahra Jawad: I think it’s really when it comes down to it, partnerships are all down to people working with people more than anything else. And to me, that’s so important. Do I trust this partner? Do I trust that we’ve got the same vision and the same enthusiasm, the same research ethics, maybe the same drive and do we think that we can work together in being able to achieve? My big vision is that every antibody is gonna need a Creasallis technology in it in the field of oncology. And I think if we share those values, then that makes a great partnership more than any monetary or name or anything else that comes. Comes with that.

Grant Belgard: How do you decide which scientific problems are worth turning into company strategy?

Zahra Jawad: I think for me like that, that it was that firsthand patient experience. It’s whatever you, whatever scientific problem turns into a company, I think it needs to hit personally for that founder. Something that you’ve seen done, been affected by close member have been affected by, because. You need something to get you through those very dark days of when times are tough. Definitely that employee that was in my team then gets me through those very dark days when I think, why am I doing this? And there’s no point, that there’s actually people who, if firsthand I’ve seen that desperate for some improvement in the solutions. So I think there needs to be some sort of personal connection to the problem. It’s not just a scientific problem. because you need that resilience to continue all the way to the end.

Grant Belgard: Can you give an example of a low in, in building Creasallis and how you overcame that?

Zahra Jawad: Oh definitely like closing financing rounds is stressful. Very low mark points. Financing, generally, raising money in general is very demotivating. You’ve got maybe 200 nos and maybe five yeses, so it’s a lot of nos that you’re listening to a lot of people criticizing you what you’re doing. Can’t see the value in what you’re doing. You haven’t managed to convince them. Then that sort of fear of are you able to finance the company moving forward? Especially when you’ve got employees, there’s a sense of responsibility to them. Okay, I took the gamble, I’m the founder. I’m the crazy one with a crazy idea, but I convinced people to join me. And those people, I have an obligation to sure they still have a salary going. And so that’s quite a low point I think in a startup generally, very small lows are hugely exaggerated and very small highs are hugely exaggerated.

Zahra Jawad: So just seeing a band on a gel and you’re like doing cartwheels down the corridor to the lab you can’t believe your luck. In hindsight, for most researchers, that’s nothing. But it’s those little wins and even things getting you down of like. Oh our tax returns wrong. We’re gonna do that again. That can make you feel really low and it’s really insignificant ’cause you haven’t been fined and you haven’t been told off. But it can really make you feel low.

Grant Belgard: How do you think about balancing building a platform and advancing specific programs?

Zahra Jawad: So we were a platform, ’cause we’ve got a general technology of how to improve antibodies and I thought initially everyone is gonna want that. It’s gonna be a platform that every pharma company is gonna want in there. Best assets quickly. And then the and it was probably like that maybe about five years ago. And then the environment changed. And what changed is that, a lot of big blockbuster drugs for a lot of these pharma companies are coming off patents and they need to fill that gap. They are looking, then they start, they’ve shifted from looking at early stage platform things to, we actually need to plug in those, we need to get assets to plug it in. And so you have to react to that. If you can keep building a platform and you can hope that will come back, or you can, you have to, keep your pulse on the market and see what your customer really wants and they want assets.

Zahra Jawad: So we pivoted about just over a year ago we decided to go for assets as well as the platform. We still do about 20% of our work platform development, about 80% on the assets itself. because I still feel that even if we wanted to pursue the platform, we’ve got to show an end-to-end process and what best to show it than on an asset that could also be partnered or developed ourselves. Um. Tangible. So we still, any partner is gonna wanna see how you go from idea to a first injection into a human being. so it’s not wasted to work on an asset when you’re a platform company, but you do have to react to your market and what your customer really is looking for as well.

Grant Belgard: What makes an idea look exciting early, but weak in the real world?

Zahra Jawad: When you haven’t done your customer discovery well at all, every founder thinks their idea is amazing. We are delusional, we are, we can’t, we’re very blinkered in our views, and I’m definitely guilty of that. And there’s nothing more grounding than talking to your potential customers and listening, not pitching listening to what their pain points are and what they really need. And if that doesn’t align with the product you’re developing, That is gonna be a weak idea. You’ve got to pressure test your idea very as early as possible, and actually not just once. You’ve gotta continuously, and it’s really on the founder to do that because you are delusional until you actually are humbled by what your, by what your customer tells you. And if your customer is not interested in your product, then.

Zahra Jawad: You’ve got, you either continue building something that has no market value, which is then a weak idea, or you have to pivot quite rapidly to be able to meet that customer’s needs. and that’s what makes a great company from one that fails eventually, in my view.

Grant Belgard: What kinds of evidence give you the most confidence that something is really moving in the right direction?

Zahra Jawad: Definitely like customer interest. I’m gonna put the customer first. ’cause I think that’s what it is. Whoever your customer is, if they’re willing to pay for you what you’re doing or partner with you or buy what you’re doing, that’s really great evidence early on that you are on the right path and you’re building something that somebody wants and values. Yeah.

Grant Belgard: What do people outside drug development tend to underestimate about getting from a good idea to a real therapy?

Zahra Jawad: Drug discovery, drug development is a very structured process and it’s really well mapped out in my view. Once you’ve, there are nuances, of course. Every project has slightly different needs depending on who you’re trying to treat and what your competition is and how you’re gonna stand out from your competitors. I think that’s one thing that people don’t understand is that you start, you have to start a project with thinking about what does your final drug look like for it to be competitive. And of course you’ve done your customer discovery and you know that this is what my customer needs, whether it’s the patient, the hospital, the oncologist, or the pharma company. This is what they need. And so you start with that sort of. Guideline firmly at the beginning of every meet, every project meeting of this is what we are looking for. It needs to bind this, but not this.

Zahra Jawad: It needs to have this kind of affinity, not that it needs to be formulated like this so that we can give it in a syringe versus giving it in an IV bag. Like all the things that would make you competitive, you need to have there in every meeting you’ve got to. Prove it or disprove that you’ve met those criteria. And what would you do if things, and of course you’re not gonna meet everything ’cause it’s a wishlist. And then it’s what are you willing to compromise on? And every drug is a compromise. There’s no perfect drug out there. everything binds something that is not meant to. Everything is slightly toxic more than we want it to be. it’s that, and that’s where the experience comes in of people going, I think we can compromise on this, but we can never compromise on that. And, And making that judgment call to take that drug all the way to market.

Grant Belgard: When you look back, what were the biggest turning points that shaped your own path?

Zahra Jawad: I guess the first my, to being a founder or a scientist?

Grant Belgard: Both. One leads to the other, right?

Zahra Jawad: Okay. So definitely scientist, I’d say people had the most amazing from my science teacher at school from when I was 12 onwards. People who encouraged me and, um got me to actually where I am, even when I didn’t believe in myself. I never thought I was gonna ever be a scientist at 12 if it wasn’t for my science teacher who encouraged me. I never thought I would ever start a company if it wasn’t for a line manager. I had, a Janice, who told me that I could go all the way. Um, and it’s always these like key mentors in my life who have pushed me and have seen something in me that I’ve often not seen in myself. And then opportunities, right? Everyone gets to turning points in their life of which you feel doom and despair. You’ve taken the wrong job. Your job isn’t working out and you are really not enjoying it.

Zahra Jawad: your contract, your postdoc contract’s coming to an end and you need another postdoc, but no one’s giving you a job. How do you react to that? those are the, usually the biggest turning points is where I’ve gone. Oh my God, I can’t get a job. I’m just gonna have to apply for everything. And sometimes life takes you to places that you didn’t quite expect and what, the market sees you is not quite how you see yourself. those are quite big turning points where was pretty sure after I finished a PhD I was gonna be an academic then until I couldn’t get a fellowship and, but industry offered me a job. I was like, oh, that’s interesting. I never thought of industry and things like that. And, it was people, opportunities that come your way when you’re at your lowest point. So I think that’s my big advice is when time, when times are tough, look for the opportunity that’s available at that time.

Grant Belgard: What drew you into antibody engineering in the first place?

Zahra Jawad: So I’ve always been fascinated by protein engineering. I think it was the one lecture at university where I would just sit there with a pen down, just like listening intently. It was so fascinating that changes in amino, one amino acid could have such a big impact on a function. So definitely wanted to do a PhD in protein engineering. Um, again, I was actually adamant I was gonna stay in protein engineering and have an academic career I know I was on my second postdoc and struggled to find a job, really to find a third postdoc or a third fellowship, or someone who would take me on. it was an antibody engineering company that they offered me a job and that was the one industry job I applied for amongst about 20 to 30 postdocs that I applied for at that time. So life, the world had something to tell me, so I fell into it.

Zahra Jawad: Still interested in engineering proteins, but antibodies I’d never thought of before. And it was there at my first sort of industry job. I was just surrounded by some of the best people in the field, um, that I was just so hungry to learn from them. That got me really into antibody engineering.

Grant Belgard: What have different kinds of organizations taught you about how science gets done?

Zahra Jawad: So there’s of course, the startup world, the mid-size, and the pharma, I’ve seen it all and I’ve worked in it all, and science gets done really differently in all of them. I’m telling you it’s drastic and I think you are. Either you sit in one of those buckets, but I don’t think you can find yourself enjoying all three environments. In big pharma, science is compartmentalized. You play one very small role that you do every day very well. And part of a bigger field. Nobody wants to know your opinion unless it’s asked for. And if you want something to change, you have to wait for governance, you have to take it to a meeting. It needs to be approved by someone senior, and then you’re allowed to change. So there’s very little freedom, but that works really well for people who love the process and the pre predictability of their job every day.

Zahra Jawad: And there are people that love working in pharma in. You can see in that kind of way, the science is done. It stifles innovation because everyone wants to do it. The process is more important than the thought process, let’s put it that way. Startup is completely different. Science is done very differently, a little bit chaotically especially, and also in universities a lot more. In a university you would do everything from start to finish to get that project completed. In a startup, it’s a bit more compartmentalized where you are not maybe as tightly compartmentalize as in a pharma company and you have a little bit of creativity, but there is some order still. You can’t just take the project where you want to. You really have to follow and get to that big next big inflection point for the company. And that’s really important there.

Zahra Jawad: Depending on where you sit and mid-size pharma is something in between where there is a lot of structure, some governance, a little bit of freedom. But, startup is because you can follow the science, there’s a lot more innovation. Pharma are not brilliant on innovation, but really good at getting everything out on time, and cost effectively as well because they are following the process. So it’s really different how science is done in different kinds of organizations.

Grant Belgard: What have you learned from roles that stretched you in unexpected ways?

Zahra Jawad: I think it’s really important to just listen. You are not gonna hear things you like and that you may be a bit hard to take. Those are the most challenging jobs where you’ve been told that you are not doing a great job, people don’t like what you’re doing. You’ve really, advice is not to react, it’s just to listen. It’s your choice. If you want to take that feedback on whether you believe it, we think there’s some truth in it, or it’s completely not true, but someone said it, there’s some truth in it, and so you need to listen. and Those are, those difficult moments. I really, the, that pain is, I really believe that is growth. That’s you growing and maturing and adapting. And, still doing that every day. I’m not, I’m nowhere near ready have nailed it all.

Zahra Jawad: There’s challenging moments all the time, whether you’re cha challenging employees, challenging investors, challenging board, challenging partnerships. Eventually you’re gonna have to learn to deal with people and the quicker you figure that out and the you stick it through and the resilience, those are the real turning points in your personality that’s worth having.

Grant Belgard: What skills have you found transfer to leadership more naturally than you expected? And what have you found doesn’t transfer well?

Zahra Jawad: Yeah, I think like I’ve always been a good storyteller and I think being a leader, you need to tell a good story or at least explain things very well and set a vision. Those, that ability from early in my career of defending a Viva, doing a presentation, being able to like even talk about the problems you’re having in your experiment and getting that sort of constructive feedback back quite similar to being a leader. You’ve still got to set that, be able to have that clear communication with the team. Unexpected things. Are different to you and I think that’s the most unexpected thing. What motivates me, what gets me up is not what gets everyone else in the team up and that’s difficult to figure out that individual, what motivates them and how can I get the best out of them because it’s definitely not what gets you outta bed in the morning.

Zahra Jawad: And that, that took me a while to figure out that people are just not motivated in the same way that I’m motivated.

Grant Belgard: What skills have you had to learn the hard way?

Zahra Jawad: Diplomacy, like when somebody is having a bad day, any, anyone, whether it’s a supplier a partner, an investor, a team member. They’re, and they’re just going at you, you’re just tempted to go back at them. So I’ve learnt that the hard way, it’s better to just be diplomatic and step back and wait for a moment where everyone’s calmed down to. Let’s talk about that again in a calm and grounded way. And, you have to practice that. ‘Cause you can feel your blood pressure rising when someone’s going at you. I do have to like practice. Internally, counting, breathing, feeling the hem of my shirt to ground myself. It all comes like I’ve had to actively, ground myself to be able to, yeah, to be able to not react when someone’s going. You, especially as a leader, you can’t react.

Grant Belgard: How has your relationship with risk changed over time?

Zahra Jawad: I, I always see myself as the most risk averse person that ever existed. So I think it’s, it was a surprise to a lot of people when I decided I was gonna go and do the riskiest thing you could do is start your own company. Not only that, start it without a shred of data to support the idea at all. So I guess. Like I said, sometimes I just don’t know myself and how much I see myself in a particular way. And I see myself as a very risk averse person. I like the structure, I like being part of a team. And then to go and do something silly start your own startup and take all that risk. And I didn’t see it as a risk. I just tell myself a different story of this is an important problem that needs to be solved.

Zahra Jawad: I need a safe space to solve it, it’s not a failure if it doesn’t work, on me, but I’m glad I gave it a shot and that kind of was the attitude I had when going into the company that I didn’t wanna be on my deathbed with the regrets that I didn’t even try.

Grant Belgard: What advice would you give someone choosing between going deeper scientifically and going broader operationally? If they’re coming from a background as a scientist.

Zahra Jawad: First figure out yourself, what you’ll really like. If you are a scientist who likes to tinker and doesn’t like the ops operational structure, stick to the scientific route. I think do what comes naturally to you. And if you are operational and you can’t stand like the tinkering around in this lab that really bothers you, then stick to operational. I’m an unusual blend of both. and I think I could give advice to those who think they are also both. I remember being told very early on in my career that Zahra, you’ve got to choose whether you’re going down the scientific or managerial career path. And I just used to say, look, I really strongly believe in leading my team from within the laboratory, so I’m not leaving the lab. And if you think that’s a scientific role, then be it. And if you think that’s an operational, a managerial role, then so be it. But.

Zahra Jawad: I’m Zahra and I’m a square peg in a round hole. And you just have to deal with that. And I remember saying that to my manager and so they left me alone. They let me have a team. They let me still stay in the lab, and I do strongly believe in leading from within the laboratory. I think the lab is a fabulous space. There’s so much. Banter, there’s so much discussion that happens over a centrifuge that you’re turning off or a gel that you’re running. There’s just so much downtime in a lab creates real opportunities for leaders to influence or have discussion and show, show positive traits from within their, to me, more than sitting from behind a desk. So I do have that unusual skill of being able to get things done scientifically, and I enjoy, I really enjoy the science and being in the lab and I am still in the lab also being able to motivate people. But I do it from within the lab.

Grant Belgard: How should someone judge whether a role is helping them grow or quietly boxing them in?

Zahra Jawad: Listen, ask for the feedback. A lot of people are really scared of asking for feedback, in our, in Creasallis, we don’t call it feedback. We call it an encouragement and and an appreciation, because I think when you say it’s an appreciation and an encouragement, it comes from a very different place than feedback. You cannot– listen to what people say about you. You’ll always have a feeling of whether this, this position is constricting you or letting you go freely. But if you’ve got really good mentors, bosses, line managers, other people in the organization, other people that are outside of the organization, anyone you’ve met who you think, I really click with you, I’ve got respect for you, network, find those people and ask them.

Honestly: is that what you think? I think a lot of young people think that every job is boxing them in it’s not allowing them to, why? Why am I not promoted? I’ve been here a year or six months already and. I hear that a lot. Be curious about the people around you and be honest with yourself. Are you really, like promotion isn’t just about how good you are. Yeah. You’re brilliant. I know that. But promotion’s also about is there an opportunity in the company for that promotion? So be curious it’s not a failure if you’re not promoted every year. So be curious and get that feedback and and be patient. You’ve got 40 year, in some cases gonna be 50 year career ahead of you because our retirement ages are rising. You don’t have to have it all in the first five or 10 years of your career. You can pace yourself and really um, be curious about the role that you are in.

Honestly: And actually you have to find your own opportunities at the end of the day. Call yourself boxed in. But there are tons of opportunities in organizations to volunteer, to do things, to put yourself forward for things that gives you that sort of variety in the role that you really need.

Grant Belgard: What makes a manager or a mentor genuinely life changing?

Zahra Jawad: Someone who can give you really honest feedback. I can’t stress that more. Some of the most life changing managers I’ve had have been the ones who’ve made me cry the most. So yeah, cry, go to the toilets and let it all out. And then listen again and to those words because it might not be what you want to hear, but it’s definitely how people are perceiving you. people only, or people would give you real feedback if they care about you. If I didn’t care about you as a manager, I wouldn’t tell, give you any feedback. And that’s not good for you. So if you are seeing someone who likes to give you feedback, don’t shun them, I think get closer to them because they obviously, want to see you improve, change, grow. All of that is positive. Hang on to those people for dear life and don’t ever lose touch with them.

Grant Belgard: What do you wish more scientists understood about leadership?

Zahra Jawad: I think a lot of the times it’s not just about the data and the facts, it’s that you’ve gotta bring people on the journey with you. I still sometimes struggle with that, with my own team, getting people to see beyond the the logic. And a lot of scientists are like, it’s logical, therefore we should do this. And it’s more about the influencing this, coming hard at somebody with a lot of logic is not gonna turn them over. You probably have the opposite effect. They’ll probably be like you’ve just dissed all my data and I am not on your journey with you. So there’s, I think what scientists underestimate is the power of influencing. Important, that we need to learn that quickly so that we can get people on board if we think that we’re doing the right thing.

Grant Belgard: And what’s something you’ve changed your mind about in the last few years?

Zahra Jawad: I change my mind all the time as new things definitely about AI. I’ve changed my mind about that. It needs to be embraced and needs to be integrated. I’ve changed my mind. Maybe investors were quite scary when the first, before I started the company, someone I didn’t wanna talk to. I think I’ve changed my mind about that. I think they can be real. Um, pillars of support for you when you’re building your company, they’ve got the same incentive as you have. They wanna see it succeed and you can really embrace them as partners. And there’s all always little things that you change your mind about continuously. And I think it’s really good to have that open mind and and admit sometimes that you are wrong.

Grant Belgard: What advice would you give to someone who thinks they may want to found a company one day, but they don’t feel ready yet?

Zahra Jawad: Don’t do it until you’re ready. I completely agree. You’ve gotta be ready. It’s it’s it’s con, it’s so life consuming. And it’s fun For the first six months, you’re like, yeah, look at me working all hours. I’m only getting four hours sleep. All I think about is the company. But you remember, it’s not a six month stint. This is could be a 10 year stint for you. So you’ve gotta be really ready that you want to do this. Being a founder and selling a company is not for everyone. It’s. Your relationships will suffer with people personally. Children, if you’ve got that families you’ll have to sacrifice. Forget taking a holiday. You will have to take a call at some point on vacation. Or you’ll have to pay a bill or you’ll have something urgent to sort out. So it’s not for everyone. ’cause some people just want to spend time. With their families and they want to get away from it all.

Zahra Jawad: And a founder is not gonna, you’re not gonna have that. So you’ve gotta be sure that you’re really ready for it before you go. So it’s fine and you don’t have to do it. Now if you’ve got a great idea and you think maybe this isn’t the right time of my life, I’ve got. A young family, I’m about to buy a house. My parents are unwell. I have to look after them. That’s totally fine too, to wait until the time is right for you because when it does start, it is o over consuming.

Grant Belgard: Where do data computation or bioinformatics create the most leverage in therapeutic discovery today?

Zahra Jawad: I think it hasn’t happened yet, but I really think it’s that problem of going from the lab into going into mice, which I think we are okay with, but it’s then going into human, going from mice to human. I really think that is where computational biology is going to help solve that, that problem for us. All that genetic information that we are collecting. Sample information. RNASeq data from patients from tumors. We need that computational power to make sense of whether my drug is going to react in that patient or not. And then to like really simplify that down to a profile of where your drug is gonna work, rather than what we’re doing now is just, give it to everyone and then we are like, wonder why we’ve only got a 20% response rate on it. So that’s where I’ve got high hopes for the computational world that they’re gonna solve that problem of data from a mouse into data into a human.

Grant Belgard: What’s one belief you hold more strongly now than you did five years ago?

Zahra Jawad: That’s a very difficult question. I think like [plugging] through in resilience is something that I hold now much stronger than before. Like not giving up, keep going. Yeah, when things are tough sometimes you just wanna go. Okay. I can see that definitely with jobs I’ve had in the past. It’s tough, bad boss, bad colleagues, I’m not able to do the research I want to, you’ve just canned my project. It was such a good project, why? I’m leaving. And you could just throw the towel in and go to another job because you’re just annoyed. And I don’t think that anymore. I think like sticking it through has got a lot of value right to the end.

Grant Belgard: And what advice do you return to personally when things get difficult?

Zahra Jawad: So I’ve got a toolbox. When things are difficult, first of all I can, I know it’s happening because I can see physical changes in me. That I know this is Zahra’s not in a good place right now. And I know when Zahra’s not in a good place right now ’cause times are tough, I need to go for a run is one of them. So the team all knows, when Zahra’s in a bad mood tell her to go on a run because it’ll make me feel better. And I know that make always makes me feel better. There’s a lot of. Like I said, the deep breathing and the grounding myself exercises that I’ve actually trained myself as a, so that doesn’t come naturally. I have to train myself to ground myself and get myself out of that. Practice just calming myself down and getting myself out of that situation. But I’ve been really highly trained in trying to recognize. When I feel stressed, I feel it in my chest.

Zahra Jawad: I feel it difficult to breathe, and I know when those signals come on. I need to open my toolbox and go for a run, make a di–, make dinner is a big one. I do enjoy cooking and I forget just going and being creative in the kitchen and just throwing things together makes me happy and eating it makes me really happy. Seeing friends, seeing my family. Going on holiday, getting some sunshine. Okay. Sunshine is a big one. Makes me a lot happier if I’m just sat in the sun for half an hour. So I just have to open that toolbox and see what’s available. If it’s a cloudy day, then it’s definitely a run or making dinner. Yeah.

Grant Belgard: That’s nice. So what do you hope listeners remember from this conversation?

Zahra Jawad: I hope, okay, number one, if you’re gonna start a company, do your customer discovery. I cannot express that enough. Do your research and make sure your customer really likes your, wants and will pay for your idea. This is a business, not a charity. Number two. There will be lots of times when lots of times when things are very tough, you just have to plow through, find a toolbox of things that make you happy. Write them down, put them in a poster in front of like your bed so that you see them every day. remem– remind yourself of when things are tough, I’ve gotta dig into the things that make me happy to get myself as quickly as possible. Out of that, you’ve gotta get yourself out of that as quickly as possible, are the two things I would say to anyone.

Grant Belgard: Well Zahra, thank you so much for joining us. It’s been great.

Zahra Jawad: You’re very welcome. Thanks for having me, Grant.

Transcript of Episode 85: Jason Mellad

Disclaimer: Transcripts are automated and may contain errors.

​

Grant Belgard: Welcome back to The Bioinformatics CRO Podcast. Today I’m joined by Dr. Jason Mellad, co-founder and CEO of OtoImmune, as the company steps into public view. Jason’s path spans Cambridge trained scientist, biotech operator, and venture builder, and now he’s channeling that experience into a precision immunology company that publicly describes itself as founded by and for autoimmune sufferers. In this conversation, we’re going to unpack what OtoImmune is, why this problem matters, how Jason’s career led him here, and what it takes to build something genuinely useful at the intersection of biology data and patient need. Jason, welcome to the show.

Jason Mellad: Thanks, Grant. Great to be here.

Grant Belgard: So for listeners meeting OtoImmune for the first time today, what is the company?

Jason Mellad: So OtoImmune is a precision immunology company. We recognize that oncology is decades ahead of the immunology field today, in most places, in least, in Western geographies. If you have a tumor, it’s sequenced and that information is used to inform care. So the genetic mutational landscape is mapped out. You can start to stratify the patients in terms of progression of disease, what they may or may not respond to in terms of drugs, and also understanding the severity and grading, and even the origin in many cases of the tumor. In immunology, you’re lucky if you get anything more than basic testing, and that is trial by error. Try this front line of care. If that doesn’t work, we’ll try something else. We’ll try something else. It can take years for people to get diagnosed because the symptoms are quite vague and there’s just not that precision that so many sufferers need.

Jason Mellad: So OtoImmune, my co-founder, Dan and I set it up last year so that we could fix that problem, bring precision by leveraging multiomics and advanced informatics to really help drive the field forward.

Grant Belgard: Beyond the general problem. Is there a particular experience that made you feel this company needed to exist?

Jason Mellad: Yes, so I can speak personally for myself. My mother-in-law has rheumatoid arthritis and has been battling it for several years, decades. And my husband and I, we had our kids during COVID and we’re so grateful that we were able to bring them to Greece finally, when the pandemic ended. So they could meet their grandparents. And because she was on methotrexate, a really harsh chemotherapeutic drug that’s frontline for a lot of rheumatoid arthritis patients, her immune system was completely knackered. And so when she met her grandkids for the first time and gave them a big hug, she caught a nursery cold like so many of us do. But in her case, because she didn’t have the immune system, it had been suppressed by the meds. She was knocked out for weeks and it was awful.

Jason Mellad: And I will never forget the moment when she came into the living room and she was in tears and she told me either I can take this medicine to use my hands, or I can be with my grandchildren, but I can’t have both. And that was a light bulb moment for me because until that moment, autoimmunity is something I’d heard of. I hadn’t studied it, but I didn’t know anybody in my family who had, it. Turns out that a lot of people, women in particular, are suffering in silence and that is very common, over 10% of the global population and increasing. And people like my mother-in-law have to go from doctor to doctor, bouncing around from treatment to treatment, trying to get care, and it’s just not good enough. That was a turning point for me.

Grant Belgard: And why is this the right moment to talk publicly about OtoImmune?

Jason Mellad: I think it’s the right moment. Not only because we’re launching and we’ve pulled everything together with my amazing team, but it’s because we know that the pace of medicine. Precision, the ability and access to things like whole genome sequencing, AI, and all of these different tools. We’re now at the intersection where people are more aware of autoimmune conditions and how important they are, that there’s more attention from pharma, developing a pipeline of drugs to treat these chronic conditions, and we have the informatic tools available to really understand the diagnosis and stratification for care. So it’s now the perfect moment where all of these elements are intersecting.

Grant Belgard: When you use the phrase precision immunology, what does that mean in practice?

Jason Mellad: In practice, it means taking the right information longitudinally to not only diagnose and in some cases prevent, but in reality, characterize the unique immune system that each of us has. Our company name Oto comes from a Yoruba word, which means unique. And we chose that with intention because we believe each of us has a unique immune system. It’s informed by our genetics and it changes over time. And so that means that generic treatment pathways, generic bits of advice you might read online is not going to be suitable for most people. So precision is by definition required for immunology because it is something that’s so unique to us as individuals and evolves over time. You need to also have that longitudinal view as well, and we want to provide both not only snapshots of analysis, but also longitudinal joined depth analysis.

Grant Belgard: And what patients do you want to focus on first and why start there?

Jason Mellad: So we’ve mapped out a wide number of personas, and it’s something that we spend a lot of time working on. Early days, we interviewed over 5,000 patients and clinicians. We’ve all have lived experience in the team, my co-founder Dan, has diagnosed inflammatory arthritis is an autoimmune condition, and most of our team members have autoimmune or autoinflammatory conditions. But we wanted to make sure that as we are building out the platform, it was very clear that different personas had to be supported. So whether or not you have a family history and we call you a vigilant person, or if you’ve got unexplained symptoms that seem quite vague, but it could be your immune system all the way through to somebody who’s just newly received a diagnosis. Certain people who are unfortunately experiencing multiple flares, and then finally when someone is in stable remission.

Jason Mellad: So the platform we’re building, we believe, will support people at all stages of those journeys from just proactively looking after their health all the way through diagnosis and through to remission, and hopefully make each one of those transitions easier. To start off with, we’re very keen to find individuals who have undiagnosed symptoms and have been floating around in the system for some time, and for those who’ve just been recently diagnosed and are dealing with the magnitude. Of that diagnosis, which is now a lifetime change for them.

Grant Belgard: What are the core ingredients of your approach?

Jason Mellad: So at the very basic level, we have a digital health platform. So we take in information from symptom tracking, medications our individuals are using, their family history, their personal health history, metadata, height, weight, you name it, and bring all of that together to build out a solid foundation. The next level up is we like to bring in whole genome sequencing. Not everybody is able to access their genome today. We want to make that as democratized as possible. Because we feel that everyone in the future will have their whole genome sequence and they’ll be very informative for care. So that’s the solid foundation and your germline genome in particular, which doesn’t change that much over time except for somatic mutations that can occur. Your germline mutation is important as that foundation.

Jason Mellad: Then we start to layer in other omics on top, so blood biomarkers, inflammatory markers, lipid profiles, and just keep stacking and stacking, gut health profiling, and build up that case of. Real world evidence coming from wearables that are integrated. So your exercise, how much sleep you’re getting, the quality of your sleep, your metadata, genetics, blood tests, gut health, and bring it all together for what we call integrated insights.

Grant Belgard: In a noisy world of AI diagnostics and precision medicine, what do you want OtoImmune to be known for?

Jason Mellad: We want to be known for impact. I think a lot of people think on AI as a big disruptor which it as it is, and there’s a bit of a land grab where. Everyone’s saying, I’ve got AI, you’ve got AI, everyone’s got AI. We are less focused on the how and more about the what was achieved. So we are very clear on why we’re doing it. We are clear on all the tools that we’re using for how we’re going to make it happen. But what I want OtoImmune to be remembered for and known for is we actually had a positive impact in patients’ lives. We were able to get people diagnosed sooner with more accurate diagnoses and on the right treatments. We were able to help people avoid getting ill in the first place. For certain conditions, it looks like that might be possible.

Jason Mellad: We’ve helped pharma unlock their pipeline of drugs and make sure that they actually get more shots on goal that actually pass through all the way into clinic because those life-saving therapies are needed. It’s impact. Not just another biotech, not just another fundraising startup, but we really made a positive impact in patients’ lives, their families’ lives, and society.

Grant Belgard: What evidence will you need to generate to earn real trust?

Jason Mellad: I think it starts with the authenticity of the team. Because we’ve used the platform, it’s benefiting us, and so we can show to people that we understand we have lived experience, whether we’re caregivers or diagnosed. Then our focus now is to really go out and solve very key issues that affect the different personas that I outlined earlier. So are you having trouble getting your GP to pay attention to you? Let’s help with that. Are you trying to navigate which of the many. Immunosuppressant drugs, you might have to use the [DMARDs] with your clinician. How can we help support that? Are you trying to understand your risk profile given that your mother and your father both have autoimmune conditions and you want to be able to catch something early so that you can intervene early? We can help with that as well.

Jason Mellad: Are you looking to understand once you’ve had your genome sequenced, if there are any variants that could influence how you would respond to different medication or your risk profile? We also would help there. So having very clear value propositions that we deliver on, and getting that positive testimonial and feedback that we’ve had that impact is essential. That’s far more important to us than going out to millions of people in day one. We’re saying every person that we help in those first days, we absolutely have to demonstrate our value.

Grant Belgard: Which career pivots best explain how you got to this point?

Jason Mellad: Oh yes, where to begin. So I originally wanted to be a. An academic, like my father, I’m a mixture, mixture of my parents and my mother is a social worker by training and was very entrepreneurial, had multiple successful jobs. And some companies she set up were successful and others weren’t. But we learnt from that. And my father was at the same university for over 40 years and was a professor of animal genetics and the Vice Chancellor of Agricultural Research. So I, I have both in me and I wanted to be an academic like my father and I quickly learned that wasn’t the path for me. Then I went into technology transfer, working at the University of Cambridge, learned about intellectual property rights, spinning out companies, getting them funded, and the importance of translation, and that really hooked me and I thought, this is what I want to do with my life.

Jason Mellad: I transitioned from that into biotech and ended up at Cambridge Epigenetics as one of the first employees, and Cambridge Epigenetics has now become Biomodal, but that opened my eyes to multiomix. The idea that your genome is there, but epigenetics is dynamic and it gives you so much more information. That kind of qualifier. The dimmer switch on the genome opened my mind to the power of omics and the importance. And then from there, set up Start Codon, my fund with my co-founder Dan, and we learned how to grow and scale across multiple companies that we helped build and support those founders from scratch. But it really showed me that I wanted to be an operator.

Jason Mellad: That I understood how the venture raising, how bringing products to market, how biomarker discovery worked, and I wanted to bring all of that together into something I was truly passionate about, and that’s how we set up OtoImmune.

Grant Belgard: How has helping other founders shaped how you’re building this company?

Jason Mellad: We have seen that there’s a lot of advice we could give from our previous experience, but also we know what it’s like for founders today. It is more challenging to raise. The regulatory landscape has changed. AI has been the latest massive disruption to the field. We understand what it takes to secure financing, to secure partnerships with pharma and biotechs to get products to market, what it means to really demonstrate value, how to grow a team, how to manage a team. These are all things that we honed because doing it for one company is one thing, but doing it across a portfolio of 28 and helping founders with multiple different platform types was another. So it really supercharged our learning, like the 10,000 hours that Malcolm Gladwell talks about. I feel like we definitely put in our 10,000 hours.

Grant Belgard: What advice would you give people building in data rich but biologically messy spaces?

Jason Mellad: I’d say, number one is be very intentional about the type and quality of data that you’re building. I think now that AI’s becoming increasingly commoditized, everyone is saying, let’s look at legacy datasets and we’ll derive insights. I think you can to a point, but those datasets and biobanks weren’t collected necessarily with AI and the scale that we have in mind. So as you’re building these data lakes, think about rubbish in, rubbish out. You’ve gotta be really clear what are the controls? How are samples taken? How was the data developed? What was the QC and QA that was involved with doing it? And then also be clear on how much you can interpret from that data. Don’t overfit, don’t over interpret. Be very focused. I think a lot of people over promise and under deliver. That’s absolutely the wrong thing to do, and also consent is so important.

Jason Mellad: It’s hard to see everything that will happen in the future, but transparent, informed, and understood consent is essential.

Grant Belgard: If autoimmune succeeds, what will be different for the people you most want to serve three to five years from now?

Jason Mellad: I think that we can shorten dramatically the time it takes for somebody to be diagnosed and to get a proper diagnosis. I think that we can really curtail the long trial by error process. That goes for a lot of the medications that people are subjected to. I think that we might even be able to help people avoid getting ill in the first place with the right lifestyle changes and preventative measures so that they don’t progress into autoimmunity. And I think that earlier detection and earlier intervention, several trials have shown that you get much better outcomes for across the board of multiple different autoimmune conditions. Also, I think that it will be hopefully. Much more visible. There are multiple patient advocacy groups that we admire and other companies in the space that we’re trying to stand on their shoulders and support them as well.

Jason Mellad: I want everybody to be aware of autoimmunity. I don’t want anyone to suffer in silence, and I’m hoping that even within the next three years we’ll have a significant impact.

Grant Belgard: So that’s very comprehensive, very end to end. Where in that autoimmune journey are you most focused right now?

Jason Mellad: So right now it is on the detection to drive for diagnosis and to help guide people through the early stages of treatment. That’s our wedge. So if you’re stuck in that four to seven year loop of trying to get a diagnosis, we want to help unstick that. And once you get through to the next phase, we don’t want to leave you hanging. We want to say, now that you’re in a therapeutic pathway, whatever that treatment pathway may be, we still are holding your hand. And we’re gonna help you navigate both understanding what your body is doing, understanding how the meds are impacting you, and then to better work with your clinician. To get the care that you deserve. We’re not here to diagnose. We are not here to treat ourselves. We are here to increase the connection between the patient and the clinician.

Jason Mellad: Make sure that they get the best out of that relationship, and really to help streamline the shortened timelines.

Grant Belgard: What kinds of signals or data do you think matter most? If you want better answers in autoimmune disease?

Jason Mellad: I think genomics is going to be so essential. There are some isolated things that we do, like HLA typing. Pharmacogenomics is becoming increasingly adopted, but there’s so much to be understood from the whole genome coding and non-coding variants and structural variants that impact our immune system in terms of the B-cell and T-cell repertoire that we develop. And the various receptors that we have in our different innate immune cells. A wide range of things are impacted by us genetically, and I feel that we can’t really interpret the standard blood tests that we have in the clinic without having that genetic layer to the full extent that we should. Of course, complete blood counts and CRP and all these other standard biomarkers are very important and very actionable today, but there’s a limit when you don’t have the genetic underpinnings. So I think the genome piece is really missing.

Jason Mellad: Also, the longitudinality of what we do. Episodic care is necessary, but not sufficient. The immune system is being impacted by diet, sleep, stress, exposure to pathogens, and you can get dramatic shifts in a person’s immune system that go far and above. You had a test six months ago and they made you, they’ll put you on a course of treatment. So we have to find a better way for people to self-report so we can monitor that change over time, and also make it more affordable and accessible to have the right biomarkers tested, the right biomarkers, inflammatory biomarkers tested at the right time as well.

Grant Belgard: What’s so hard about building in a space like this where symptom subtypes and trajectories are so variable?

Jason Mellad: I think there’s always a cost factor and an implementation factor. We need to make sure that we empower patients to be central to the journey. We need to make sure that the logistics makes sense because it’s very rare, especially for those who are adults. With autoimmune diseases that their lives are not carrying on. At the same time they’re dealing with the condition. So we have to work around families, work schedules. People forget to get their bloods drawn so they miss their appointments so they can’t get their prescription renewed. Then they end up flaring up the, all these things about just the basics of life not coming together. We’re hoping to streamline that as well too.

Jason Mellad: Make it easier for people to record their symptoms, make it easier for them to get through their appointments, make the baseline strong and robust so then we can then go to the things that are more advanced, like the multiomic testing. But even getting the basics right is important.

Grant Belgard: What role can computation play in this and where do you think it gets overhyped?

Jason Mellad: I am very much in the vein of. Computation and AI in particular is a tool that we can leverage to make things more efficient, but it does not fully replace the human in the loop when it comes to medical care. So I think that there are a lot of organizations and companies that are trying to say. We’re a doctor in your pocket, we will replace the clinician. AI can do it better. We say the opposite. We are more about how can we leverage AI to ease the burden on patients, to help them provide the information that they need to doctors to get the best care possible, to gather the evidence they need to drive diagnosis and further testing. How can we help doctors manage their caseload? They will have a number of patients under their care, how can they better organize and address their needs? That’s where AI is powerful.

Jason Mellad: It’s a facilitator that can operate at scale and really remove a lot of the hurdles that are blocking the field, but it’s not a replacement for humans, at least in this case.

Grant Belgard: What have you deliberately chosen not to tackle yet?

Jason Mellad: We’ve deliberately chosen not to tackle diagnosis because we know that people can have multiple immune conditions, and quite frankly, the old school designations aren’t fit for purpose. I’ve met so many people who’ve said they’ve had lupus or rheumatoid arthritis with totally different disease progressions. You think. Is Lupus one disease? Is it two? Is it 15? There must be so many subclasses. So for us to get into that realm, we first want to really understand the underlying driving mechanisms behind these disorders and then start to classify people based on that. That’s very much informed by precision oncology. We’re now seeing, we’re moving away from saying breast cancer and liver cancer only. But more into what is the molecular profile of the cancer? How is it responding? What are the drivers? Because there can be breast cancers that look like liver cancer.

Jason Mellad: There could be liver cancers that look like colorectal cancer, even if they weren’t derived from that tissue to begin with, because of the molecular patterns we see in the genome and other omics. So I’d like to get to a point where immunology is in that same space where we are looking at our unique immune system, we take a detailed molecular review as well as the phenotype of the individual. And we say, okay, based upon what we can see, this is what we think is driving your phenotype, and here’s the direction of travel. As opposed to, I’m diagnosing you with lupus, or I’m diagnosing you with IBD, or I’m diagnosing you with rheumatoid arthritis. There’s still a place for that. But before we step into the diagnostic realm, we want to focus on drivers and understanding how we can cluster all the different subtypes of conditions.

Grant Belgard: Where do you think the first real foothold is for the company: research, clinical workflow, decision support, partnerships, something else.

Jason Mellad: I think the care journey and clinical decision support is going to be a real big wedge into the market for us. I think when we first started, and as you, I’m very big on informatics. That’s my background. I’m big on big data. I naively thought, oh, the multiomics and big data will just solve everything. And then after speaking to patients and clinicians and payer providers and also pharma, you started to realize, wow, there’s just a lot of logistical things that we need to sort out first before we even get there. I have always cautioned founders that we invested in at Start Codon that it’s very easy to get enamored with your hammer, and you go around telling everybody they’ve got a nail. And you’re just going around, bashing around look at my beautiful hammer. It’s like Thor’s hammer shoots lightning bolts and it’s made from platinum.

Jason Mellad: And trust me I got my multi omic hammer and they go, actually, you know what? I just need to make it easier to make an appointment. Or I just need a way to put my symptoms together so when I go to my doctor, I can have a coherent conversation in the 10 minutes I have to get their attention. Those basics. If you don’t fix that, no one’s talking about the whole genome. So we said fix that first, and then we can make sure that we unlock the bottleneck towards the multiomic future.

Grant Belgard: That makes a lot of sense. What partnerships or collaborators matter most early on for you?

Jason Mellad: I think obviously we are reaching a lot to patient advocacy groups because they have their finger on the pulse of what the communities need. We are really keen on working with clinicians and hospital care systems to get the best understanding possible of what their pain points are, and hopefully leverage them to get more patients and customers using moi. Also biotech and pharma. We understand that there are existing data sets and biobanks available. There are also CROs that do things like clinical trial recruitment and a lot of NextGen platforms with AI. But we want to be a go-to partner because we’re focused on immunology. We are building from the ground up with that end goal in mind partnerships and that we ’cause of our backgrounds working with pharma and building biotechs. We understand some of the needs.

Jason Mellad: For example, something as simple as knowing that the data you generate, also having a biobank sample so a partner can generate their own data set with their own assay is also important. I think a lot of people don’t fully appreciate that. Knowing that the metadata that comes with the sample is so important as well. How is it processed? How is it stored? Who did it come- All these things are like the details that when you, a partner comes knocking, they say, wait a minute. You didn’t have these regulations in place, you didn’t have your LIMS set up. Your QMS is a, all these are things that we’ve lived through, and so we’re building now with that partnership goal in mind.

Grant Belgard: What milestone over the next 12 to 18 months will matter most for you?

Jason Mellad: I would say market traction, but not in the sense of volume of users, but more in the testimonials that we’ve had, the impact. That is our full focus. Now. We’d much rather have a few hundred really dedicated, very intense users who we can show we are categorically changing their lives in a positive way than thinking, oh, let’s go for thousands and thousands of early adopters who aren’t engaged at all, and we’re not able to show any impact because they, inevitably there’ll be an attrition and they will drop off. Any investor worth their salt, any business person worth their salt knows that’s a false, kind of comfort to say, oh wow, look at the adoption we had early on. You need people who are truly engaged and are really feeling that you’re solving a pain point for them. So those are the milestones for us.

Jason Mellad: Get our first cohort of, a few hundred really devoted testimonial customers who can go out and say, this works.

Grant Belgard: What would a thoughtful skeptic need to see before believing the company can make a meaningful difference?

Jason Mellad: I think first and foremost, they need to talk to somebody who’s used our platform and said, okay, this isn’t a nice to have, it’s essential. I think also they’d need to know that we have the ability to scale, because starting off with a hundred people is one thing, but going from a hundred to a hundred thousand to a million plus is another. So I would say that if I were outside looking in, is this scalable or is it just a niche thing that you can only do at a very small level? So we need to demonstrate as well, scalability.

Grant Belgard: What part of your scientific training still shows up in how you operate?

Jason Mellad: Wow. I recently went back into the lab to start tinkering around with some new workflows because I have an assay development background, particularly for my days at Cambridge Epigenetics/biomodal, and that’s been a lot of fun. Took me two weeks to get a PCR working, but that’s fine. I had support. And a shout out to the wonderful people at O2H Labs in Cambridge. But that is, I think, really important for us too, that we are not just relying on off the shelf assays. We now identified the type of data sets that we want to build and the scalability and making sure you get robust data. And so we’re developing our own proprietary assays. And so since that’s my background, it’s been a lot of fun to really think through, not just a me too assay for the sake of it, but one that will generate the data sets at the level that we need affordably to drive our insights engine.

Grant Belgard: What part of company building did you only learn by doing?

Jason Mellad: How everything goes wrong, crisis management. I would say that’s it, because you know,

Grant Belgard: That’s That’s fair.

Jason Mellad: Everything you put on paper, everything that ChatGPT tells you and all that wonderful stuff, great. Okay. That’s business 101. What you learn by doing is when things go pear shaped. When you’ve got an HR issue or the market shifts, or the experiment didn’t work, or a whole host of things, a new competitor pops up, whatever it may be, you learn by doing.

Grant Belgard: How different does it feel to be in operator mode?

Jason Mellad: Oh, I love it. It’s interesting. It’s like when you are doing something that you are happy doing. Like when I was an investor I was very happy ’cause I love supporting other people. So that, that did feel really good. But now it’s wow, I’m back into what just, it just felt like getting back on a bike. It feels so good to be in operator mode again. And although I do continue to support the founders that we backed and I still continue to do a lot of community work trying to support anybody who’s going on this journey themselves. I know this is exactly where I’m meant to be.

Grant Belgard: What have you had to unlearn from earlier chapters of your career?

Jason Mellad: I’d say that I’ve had to unlearn that hard work is the name of the game. When I was growing up, I always had this message, rightly so from my parents, about work harder. You’ve gotta be a hard worker. You’ve gotta be a hard worker. I learned over the years that you have to be a smart worker. So I always say work smarter, not harder. How can you leverage your team members? How can you use different platforms to do things more efficiently? What are some things that actually, at midnight, there’s no point sending that email. In fact, you probably shouldn’t, go get some sleep. These are all things I had to learn over time that I’m far much more efficient by working smarter than thinking it’s just working harder and potentially burning out.

Grant Belgard: What kind of team are you trying to assemble around this problem at OtoImmune?

Jason Mellad: I am so happy that everybody on the team has lived experience and is really, truly passionate. That may not always be the case, but it’s something that we’re emphasizing even for the website. We don’t want to use stock photos. We want to use images of people who actually have lived conditions, whether it be a caregiver or a diagnosed immune condition. This is something that’s core to the ethos of the company for how we present ourselves and the team that we grow. People who have lived it are super passionate and will stop at nothing to fix it for others. I love that. The other thing is I want to have diversity at the heart of it. I’m a, I call myself a Jedi, so justice, equity, diversity and inclusion warriors, it’s more fun to say that. So I love being a Jedi, but it’s because diversity is the answer. It’s not a problem.

Jason Mellad: Too many people think, oh, diversity, is this a thing, and it’s all a fad and it’s all being woke and it’s, no, categorically we know more inclusive teams that are more diverse, win. Every study has shown that. So I want diversity of thought, gender, background, neurodiversity, educational path, you name it, geographical. And we’re already very diverse, and I want to maintain that because I love that creative friction where people come at it from different perspectives and different fields.

Grant Belgard: What do first time founders underestimate when they build around diagnostics, health data, or precision medicine?

Jason Mellad: Just how challenging it is to get funded. That sometimes diagnosis-

Grant Belgard: Especially right now.

Jason Mellad: Especially right now. Sometimes diagnosis is the ugly stepchild of therapeutics. And a lot of investors will say, so you’ve got this platform and you can do analytics, could you use that for drug discovery? And you go, yeah, I could, but this is my, they’re like, oh, that’s interesting. Have you thought about maybe developing an asset pipeline and going to therapeutics? And they tell this to this team of people who are wholly unsuited to go into therapeutics. It’s not their background at all. And then they try to pivot and the next thing you know, they fail. And it’s oh, too bad. I’d say a lot of founders, speak to the right investors, speak to the right partners, and build a team for what it is you want to accomplish.

Jason Mellad: Know that it’s hard and challenging, but also understand the health economics of what you’re, the problem you’re trying to address, and know that although it’s difficult doesn’t mean it’s not possible. I think that’s the thing people, they mistake things being challenging or difficult with, is it worth doing? So just because it’s harder to get funding or it’s not as sexy an indication or whatever it may be, if the health economics makes sense, you’ve got a clear care pathway and a path to market, and it’s what you’re passionate about, you will find a way, be creative. Don’t think, oh I’m just not going to do it at all. There’s always a way. You just have to rethink about your path and find the right partners and speak to the right people.

Grant Belgard: What should people validate before they get attached to a solution?

Jason Mellad: That the problem exists and also that a solution is worth building. I think the third thing is there’s a difference, and I learned this in when I was working as a consultant in early days. There’s a big difference between want state and need state. So for anybody out there who has children like I do. They may need to eat their vegetables, for example, but do they want to eat their vegetables for dinner tonight? They need to brush their teeth, but do they want to brush their teeth? Do not assume that just because you have something that people need and you spent so much time characterizing that unmet need, and you’re so clear and you’ve validated it, and you’ve reinforced yourself and you’re running off, do they want it? Will they pay for it?

Grant Belgard: Very good point. How should a scientist decide if they should found a company, join a, maybe a startup or stay in research at a, a more established organization?

Jason Mellad: Well, it’s a personal decision, but it always starts with introspection. So I tell people you have to be self-aware. It starts with them. If you don’t understand your personal motivations, if you don’t understand what it’s you’re trying to accomplish. What’s your rationale for doing this? You should be running towards something not away. It can’t be, oh, I, my PI’s crazy and I’m, I don’t wanna be a postdoc for life, so I’ll go set up a company or it also can’t be, I want to be in control. There’s a great article called The Founder’s Dilemma, where they talk about do you want to be rich or do you want to be king? King being it’s all about control. Those companies typically fail or rich, meaning everybody wants to be successful, not rich in the monetary sense. That’s part of it. But the whole team wants to be successful.

Jason Mellad: So if you’re the kind of person who wants everyone to succeed, you have a burning desire to have an impact in the world. Pick your lane. It may be a startup, it may be joining somebody else’s. It could be in academia, but you’ve gotta decide what it is that you personally want and then know that you can’t do it all. How are you going to build a team around you to pull it off? And where do you best fit? Just because it’s your idea doesn’t mean that you should be in charge of the company. Just because you’ve got an idea doesn’t mean it’s a good one worth pursuing and be open to the idea or the fact that most people pivot multiple times. So be prepared to actively, objectively pressure test your idea. Seek critical, constructive feedback early and often. I’m always second triple, quadruple guessing what I’m doing and trying to get people to tell me, no, you’re bonking mad.

Jason Mellad: Like I, that’s, I don’t look for people to convince me if I got a good idea. I look for people to convince me that I don’t. I think that’s some of the most valuable conversations ever. ’cause if you came with the idea chances are you already think you’re, oh, look at me, I’m clever. I wanna go out to people. I’m like, please explain to me why my idea makes no sense. Give me all the reasons why this won’t work. Let’s talk her through. That’s invaluable. Actively seek that feedback.

Grant Belgard: What question do you wish more people asked about autoimmune disease?

Jason Mellad: Why is it so prevalent in women? We know that for certain autoimmune conditions, it’s roughly 50/50. In some cases, it can skew a bit more towards men in a couple of them, but overall, for the almost a hundred different autoimmune conditions that have been diagnosed, 80% of women or 80% of sufferers rather are women. Some in the past thought perhaps it was due to women going to the doctor and getting diagnosed, but I don’t think that’s quite it. Maybe that a little bit might be contributing of just who’s going and getting diagnosed. I think there’s something fundamental about the biology of XX chromosomes and women that is driving this massive disparity between the numbers. And we don’t ask enough because women historically have been underserved in the medical community and the research community.

Jason Mellad: Couple of papers have come out with hypotheses, but they haven’t really been robustly proven. We need to figure it out instead of just accepting it as is.

Grant Belgard: What belief are you betting on that more people will agree with in five years?

Jason Mellad: That diagnostics do matter, that they’re important, and that early detection and prevention is the future for medicine. Prevention better than cure, early detection, early intervention. Timely detection. These are all things, I think we’re moving in that direction already, but I feel firmly we’re going to wake up one day and realize that’s the truth. And also that we as individuals are fully responsible for our health. We cannot assume that somebody else is out there is going to catch us when we fall through the cracks. So to the extent that we can control our diet, our lifestyle, educate ourselves. I think it’s so important that we’re not kept in the dark and that we have the information we need to make informed decisions with our clinicians advising us.

Grant Belgard: What would delight you most to hear from a patient, clinician or a partner someday?

Jason Mellad: What would delight me the most would be for them to say, I don’t know. And I mean that because I think often as a patient, you’re going to a doctor and you’re expecting a definitive answer. And as a doctor, you’re hoping that there’s gonna be some clear diagnosis, some clear treatment pathway, maybe follow the guidelines, maybe go a little off rails. But there’s something very definitive and patients want definitive answers. And then when it’s a misdiagnosis or a missed diagnosis, people get really upset, rightly and get really frustrated. I would love for us to be able to have open and honest conversations where people say, I don’t know what’s causing this, but we’re going to figure it out together. That would make me joyful.

Jason Mellad: It, I think it would mean that there’s a cultural shift in the way that we do medicine from learned doctor, ignorant patient going and getting told what’s what to hey, we’re a care team working together and together we’re going to figure out what’s going on, and we need this data input from you. So you’re playing your part and we’ve got our capabilities. So we’re playing our part and together we’re going to sort it out.

Grant Belgard: If we have you back in three years, what do you hope we’re talking about?

Jason Mellad: Oh, I hope we’re talking in three years about the number of precision medicine opportunities that are now mainstream thanks to AI and the markets opening up and really backing companies like OtoImmune and others. We obviously are focused on what we’re doing, but there are so many great ideas and amazing teams out there who need support as well. And I would love to come back and talk about all the really cool things that I, we see in the market and all the amazing advances that are happening. So not just OtoImmune doing well, but everybody, ’cause I believe a rising tide lifts all ships.

Grant Belgard: Jason, thanks so much for taking us inside OtoImmune. It’s such an important moment. I like that this conversation gave us both the company story and the founder story, what you’re building, why you’re building it, and how you’re thinking about doing it. Thanks for joining us on the podcast.

Jason Mellad: Thanks for having me.

Outro: To learn more about mOI and immune health, visit moihealth.com.

Transcript of Episode 84: Eric Green 

Disclaimer: Transcripts are automated and may contain errors.

​

Grant Belgard: Today on the Bioinformatics CRO Podcast, I’m joined by Dr. Eric Green, chief Medical Officer at Illumina. Eric has spent decades at the center of Modern Genomics with a career that spans the Human Genome Project more than three decades at NHGRI and 15 years leading that institute. As director at Illumina, he now helps guide global efforts around clinical genomics, access to precision medicine, and the diversity of genomic data.

Grant Belgard: We’re going to talk about what he’s doing in this chapter, how his path unfolded, and what advice he has for people building at the intersection of biology, computation, and medicine. Eric, welcome.

Eric Green: Thanks. It’s great to be here and it’s wonderful to talk to you again.

Grant Belgard: We’re very excited to have you. So how would you describe what you’re doing now, both day to day and at a high level?

Eric Green: Let’s start at a high level. To say that I’m only completing my third month here as not only as Illumina’s Chief Medical Officer, but it’s only my third month working in the private sector. I have a long career. First started in academia and then, nearly 32 years in the government. Now only three months in, in not only the private sector, but then at, specifically at Illumina. At a high level Illumina is absolutely at the forefront of efforts to try to see genomics have maximal positive impact on the practice of medicine. Something fundamental to how I got in genomics in the first place and certainly a centerpiece of my efforts while at the National Human Genome Research Institute, in particular when I was the director for 15 years.

Eric Green: And so when they approached me, which really wasn’t that many months ago, after I retired from federal service a year ago listening to their ideas about how they wanna morph themselves from being a very technology oriented company to being a solutions application oriented company, and particularly in the arena of genomic medicine implementation, was like music to my ears. And and so at a very high level I now get to the honor of serving as a chief medical officer of what I think is the premier genomics company that really offers the greatest promise. To really push the envelope by bringing their technology, their expertise into an arena that is primed and ripe and ready to see genomics truly make an impact on patient’s life and improve human health. that’s, that is a very high level statement.

Eric Green: I will also tell you, which is what, the day-to-day part is more difficult, is the other really wonderful part of this job is um, I’m gonna craft it. Illumina has had Chief Medical Officers in the past, but they were at different times. It was, you know, you know, a while ago was their first, and then the last one that was here, hasn’t been here for three years.

Eric Green: And it’s just a different moment in time for so many reasons. And so one of the other great things about this opportunity is really being given a big block of wet, clay, that I’ve now started to mold it to see what the things are that I think we should immediately be working on.

Eric Green: But I am quite certain. Even six months from now when maybe I have a better day to day understanding of what I do, I bet you two years from now, it’ll even be different than that. Because I, I think there’s a very fast dy, fast changing, very dynamic area and to, to, to some extent. I may have ideas, but the more I learn about Illumina, I might even change my approach to what I suggest that they do or how I help them formulate their strategic priorities in terms of genomic medicine implementation.

Grant Belgard: What does the field look like from your seat right now that most people don’t see from theirs?

Eric Green: If I can tell you that 16 years ago, which is I think about the time that I became the director of NHGRI, and the big push I made as becoming the new director was that I thought, and this was around 2010, that I thought the time was right to get serious about investing in research to guide the way on how genomics is gonna be used in the practice of medicine. And we did this in a fashion where we weren’t being overly scripted and saying, all right, first we’re gonna integrate it here and then we’re gonna integrate it there, but rather we, you know, let a bunch of flowers bloom. And invest where you can in terms of research dollars from the NIH and see what happens.

Eric Green: But there was a lot of skepticism. If you go back 16 years, I can tell you that not everybody thought the time was right. The big difference now versus then is that, we proved that absolutely the time was right and starting, eight years ago, five years, depending what area you’re talking about, nobody questions the fact now that cancer genomics is real. That the use of genomics as part of cancer diagnostics, as part of cancer, therapeutic development, et cetera it’s just, it’s absolutely gonna be more and more mainstream. It’s still early days, but nobody questions. It’s gonna be real.

Eric Green: Nobody questions anymore. The idea that genomics is now a frontline diagnostic tool in helping patients who have rare genetic diseases. It’s just routine and in fact, it’s finding more and more, niche uses in that area, including in acutely ill settings such as intensive care units where especially in the neonatal period and even the pediatric period, there’s even some evidence that might be relevant in the adult ICU setting.

Eric Green: Where, patients have come in with undiagnosed rare diseases. If you could rapidly sequence their genome, you can change their medical management. This, nobody’s, nobody would’ve been convinced that was gonna be the case 15 years ago. That’s absolutely mainstream now. Or take an area like prenatal genetic testing. Where now non-invasive methods are used millions and millions of times for pregnant individuals around the world. And these tests are not very expensive. They’re increasingly being covered by insurance and it’s now mainstream that you offer to pregnant individuals non-invasive genetic testing using very powerful new technologies, some developed and refined by Illumina and others that have built on those same kind of technologies and adapted them in other ways. So these are just the early examples. But there’s more coming and there’s a lot of excitement around pharmacogenomics.

Eric Green: There’s a lot of excitement. Even around prevention and thinking about more and more that we can learn things and maybe we learn it earlier and earlier in life, that may help how patients are medically managed, which is why what I’ve gotten very excited about over the last two to three years is the seriousness that many countries are taking in studying the idea of sequencing the genomes of all newborns. To expand upon the well established genetic screening that goes on in most countries for some modest number of rare genetic diseases. And building on that to say, wow, why are we limiting ourselves to just, a handful of diseases? Why don’t we just get comprehensive genomic blueprint information that we could learn a, we can action around and even more than what we’re currently doing.

Eric Green: And yet we know that in the coming decades we’ll even figure out more and more ways to make these things actionable. And so there’s lots of studies multiple ones in the US and dozens around the world.

Grant Belgard: So you gave several examples of incredible applications of genomics that are being rolled out at scale now. What are some areas of genomics that have stayed stubbornly hard. Where maybe they, they didn’t meet people’s expectations from some years ago.

Eric Green: The most obvious area to go to stubbornly well, there’s two kinds of stubbornly hard actually. I mean, maybe that’s the first thing we should do is if you say, what are the stubbornly hard barriers to genomic medicine implementation? I would immediately bifurcate and tell you two different bins.

Eric Green: One is biological and medical complexity, and two is society and the complexities of healthcare. I think we always should be separating those. Not that one is easy and one is hard. It is just that we should recognize some of this is because biology is really complicated and physiology and pathophysiology is really complicated and unraveling that, yes, technically we can generate immense amounts of data and about things that hopefully will one day give us biological insights about it.

Eric Green: But getting those insights is still hard and then we need to separate that from health, healthcare is just so complicated in every country in the world. And we are now in a circumstance where we are trying to bring a significant change to healthcare by, in the use of genomic information.

Eric Green: And we’re running into all of the usual challenges that come with any change in healthcare. So let’s, we, if you want, we can talk about both separately, but let me focus on the biological complexities. The fact is if I, if we go back 23 years when the Human Genome project ended. We, we knew we had a challenge ahead of us in terms of actually interpreting the 3 billion letters roughly, that we had ordered. And we knew it was gonna take probably multiple generations to fully interpret it. But I think a lot of the thought at the, at that time was, all right, we’ll get this inventory of the many tens of thousands of genes we have.

Eric Green: And then when we sequence this and lots of people we’ll find lots of mutations in the genes that’ll tell us a lot about human disease. And, what’s transpired in the last 23 years? First of all, we learned that we only have 20,000 genes. Not that many more than worms and flies. So human complexity does not come in our gene count. And then we learn that holy cow, there’s all this non-coding DNA. which ha is many cases, in many areas, incredibly conserved through evolution, which was immediately signaling to us. That means it’s biologically important. ’cause otherwise evolution would’ve gotten rid of it. And it’s in, in a, the part of the genome that we don’t have the knowledge of the language.

Eric Green: We, we don’t really know how to read out, like the way we can read out the code on table to understand the coding parts of the genome. And then the more we keep studying these things, we also realize, wait a second, there’s an awful lot of reasons to believe that genomic variants are living out in these non-coding regions. That, and that these are the genomic variants that when we do these fancy human genetic studies are the ones that seem to be statistically associated with some of the most common diseases that we need to completely tackle and understand and eventually be able to make predictions about, make better treatments for, diabetes and hypertension and cardiovascular disease and so forth. And one, the, one of the big challenges that remain is learning. About the biology and the medical consequences of variation that exists outside of coding regions.

Eric Green: And there’s every reason to believe that there’s an immense amount of medically important, hopefully medically actionable variation out there.

Eric Green: But we’ve gotta define it. We’ve gotta understand it. We’ve gotta mechanistically characterize it, and that’s really hard. And we’re getting better at it. There’s lots of exciting things that are being done. And yeah, at some point you may wanna ask me about AI ’cause it seems that everybody will always ask the questions about no matter what the topic and in the world is these days, you always say, how is AI gonna influence it?

Eric Green: And of course, I do think AI is gonna have a considerable influence on helping our human brains understand all the intricacies of genome function and, and once again, I wanna stress that over the last 23 years since the genome project ended , as we realize that our complexity, a lot of it lies outside of our coding regions. Thank goodness we’ve had brilliant scientists discover all the different ways that RNA is biologically important and all the different subtypes of RNA and all the craziness of, we’re learning more and more about RNA modification we already knew about RNA splicing and alternate splice forms, et cetera, et cetera. Oh, and of course then there’s epigenomics and thank goodness that the technologies for sequencing DNA have become increasingly adaptable to study epigenomic modifications and. There we go. There we are.

Eric Green: So, you know, all these technologies are just lining up really nice. Oh, and the other variable, of course if we’re sort of saying what are the big barriers, is our, I know know what I learned in medical school. I, I learned about the kidney and I, yeah, I learned about sub parts of the kidney. But the fact of the matter is, even the sub components of the kidney, we label ’em with one name, but of course they have multiple cells within it, and not every one of those cells is behaving the same. And so there’s just a lot of biology at the single cell level. And fortunately we’re getting technologies to be able to study, all these complicated things about omics as we would often say, and doing it at a single cell level. There’s a lot of great science to be done and I and I will say, and maybe this is an important point to, to really make, is.

Eric Green: I do think sometimes when people talk about precision medicine and genomic medicine and multi omics of human health I think sometimes people think because, and they hear it from people like me and others who are very excited about it and we love to point out the home runs ’cause we love to demonstrate that, wow, this really can change clinical management. But I think occasionally people might, and especially younger people, trainees might get the impression that everything proximal to that has been solved. You know that, oh, now all the emphasis is on patient care and patient implementation. And I, every once in a while, just like to fully admit that everything we are doing that has gotten us there is cream skimming. We are, we’re just skimming the cream to be able to show how we can impact patient care.

Eric Green: But left behind in the wake is still the vast majority of fundamental biological knowledge that we need the next generation to pick up the mantle. Some that are gonna be doing overwhelmingly basic research that you know will be a bazillion miles away from clinical care.

Eric Green: And that’s okay. We need them because there’s every little step along the way that we have left unanswered questions. And one, one of the messages I would always like to give and I even think it applies to people doing bioinformatics data science, AI, is, you could say that’s what I’m doing, but you could apply that at the most fundamental level of how a genome works or how genomic variation influences genome function. Or you could use a, [?] to really help implement genomic medicine. But that full spectrum of activities will continue, I’m sure, for decades to come. And anybody who wants to get into that should feel very comfortable to know that there will be opportunities across that entire spectrum that doors are not getting shut.

Eric Green: We are just simply, just trying to open as many as we can and there’s so much more knowledge to be captured.

Grant Belgard: Where do you see the biggest gap between what genomics can do today and what the healthcare system’s able to absorb?

Eric Green: Yeah, so that’s a, it is a, it’s a great question. And you, the answer is there’s just lists and lists of reasons. What I’m learning and maybe even learning a little bit more in the most recent three months at a company that is global compared to my previous 31 years in a federal government that mostly focused on the US.

Eric Green: We, there’s a lot of interest in NIH about how healthcare is delivered worldwide, just getting my head around US healthcare was complicated enough. I, but whereas Illumina thinks more globally, there, there’s just, every possible complexity that you bump up against. And it’s different at every country, but one of the things I would like to point out, ’cause sometimes people say, ah, genomics, just, it’s so disappointing is you’re just not getting that clinical uptake. There is all this promise. I think, again, part of the problem is. Is you’ve gotta take a step back and just fully appreciate how new and how young all of this is.

Eric Green: And one of the points I always like to make is the word genomics didn’t even exist until 1987, which is the year I graduated medical school, and I say that because most of my medical school classmates are still practicing medicine and I got news for you. They never heard the word genomics once in medical school, ’cause the word didn’t exist. Okay? So you gotta feel a little– now, now I’ve gone off and become a genomics expert; they’ve been busy, treating patients and taking– and all of a sudden we thrust onto them all this new abilities in genomics. And, practicing medicine is really hard and they’ve got a lot of stresses and strains. They’re constantly fighting with insurance companies. They’re constantly dealing with a lot of regulatory issues and constantly dealing with a lot of hard problems. Then all of a sudden you throw this genomics thing.

Eric Green: I mean, we haven’t quite gotten to a point where this can be push button for them, and I’m very sympathetic, but why are we there? We’re, there, we’re victims of our own success. If genomics would just slow down for a while, then we’d have time to train the next generation.

Eric Green: But we’ve been so successful that we’ve now thrust upon the current medical system, a whole new set of things where they’ve never had any formal training. And oh, and meanwhile I would also, they’re also encountering, patients who either don’t understand anything about genomics; they, you have patients who actually know quite a bit because they’ve had time to to chat GPT it or to Google it. So sometimes, and I hear this from some of my medical classmates who I talk to they walk into a patient’s room. The patient’s already had their genome sequenced, they’ve already had all these entire, they’ve had time to do it. And the, my poor classmate never had any training on how to deal with the data, the patient’s way ahead of them. So we’re, we have a little bit of catch up to do.

Eric Green: It’s, some of the things I’m gonna be thinking about as Chief Medical Officer here is what do we do to think about genomic literacy enhancement for patients, for healthcare professionals? It’s, and it’s not just physicians. We gotta think about pharmacists and think about pharmacogenomics becoming so important. I think pharmacists can really be helpful here. Nurses, physician’s assistants, genetic counselors, so on and so forth. I think there’s that. I, then, the other example I would give is and this is where I think it’s both, it’s both a challenge now, but it’s also gonna be part of the solution is you don’t have to go– well, well, let’s just go back to the genome project completion 23 years ago.

Eric Green: 23 years ago, there were not really electronic medical records, at least not in the United States. Okay. And now everything is going to an electronic medical record. Now all of this, a lot the way medicine is practiced is interfacing with electronic health record systems but we have not been able to yet have seamless integration between genomic testing, genomic analysis and then get it all fed into an elect, electronic medical records and have it then be with immediate recommendations and all these things called clinical support tools and so on and so forth. Now, the, so the challenge is that we haven’t gotten us all together ’cause it’s all happening so fast.

Eric Green: But the good news is, is actually, these are all sort of made for one another. Genomic information is digital information. More and more medical information is digital information. More and more physicians rely on clinical decision support with popups and assist and blah, blah, blah. And so when we get to the point where we can get all this design correctly and if only we had time to build the airplane and then fly it, but we don’t because we gotta build it while we’re still practicing medicine. We’ll eventually get to that point. But what it does require is oftentimes health systems investing. The infrastructure to get the genomic information, to get it integrated in the health records, et cetera, et cetera.

Eric Green: And it’s not necessarily gonna generate money, generate revenue. It’s gonna be, you gotta make the investment in the short term to get the long term return. And some health systems are surging ahead and they believe it and some are not. And there, there’s just a lot of challenges associated with that.

Eric Green: And then, meanwhile you gotta deal the whole issue about reimbursement, who’s paying for this, et cetera, et cetera. So again, what is more complicated in society than healthcare. And yet what we’re trying to do is really change healthcare. We are, it’s not surprising that we’re seeing a lot of challenges come up.

Grant Belgard: What does good judgment look like in a field that’s moving this fast?

Eric Green: It depends who you ask. And and it depends where you sit, right? And it is it is interesting. And it even speaks to issues about, how when you have, nothing’s unlimited and certainly not, healthcare dollars are not unlimited. There’s lots of decisions that get made all the time of how do you distribute resources across an area and I, and the thing about genomics is that that there’s a lot of immediate utility, but there’s a lot more long-term promise. That, those of us like me really believe in this, but we know it’s gonna be a lot of work and we don’t know exactly the time. And so then what does good judgment look like?

Eric Green: People who are running health systems, people who are making these decisions are gonna have to make actuarial decisions about what problems they’re trying to solve that are here and now problems versus. As the famous Gretzky quote, skating to where the puck is gonna be, making the investment because you truly believe that you’re gonna be in a different place, that if a little bit of financial pain now will put you in a better position down the road and different, I can tell you different countries are tackling this differently for the same reason.

Eric Green: The different countries have different approaches to how they feel the government’s responsibility is with respect to providing healthcare to their citizens. Um, United States is not leading in genomic medicine implementation. That’s just a clear fact. And a few other countries are the, the UK in particular. When I was at the National Institutes of Health, I really, we were watching and interacting and really admiring what some other countries were doing, in part because they were experiments, right? All of us are a heart of, I’m a physician scientist. I’m an experimentalist. it was interesting to watch what other countries are doing. Now at Illumina I have a slightly different set of responsibilities.

Eric Green: And because we’re global, I mean we’re partnering with some of these other countries and their health systems and trying to figure out how we could be helpful and how can we collaborate and how can we facilitate the work they wanna do.

Eric Green: Because long-term goal is to facilitate genomic medicine and precision medicine implementation worldwide.

Grant Belgard: What first pulled you into this field?

Eric Green: That’s a great question.

Eric Green: So, So my story in the field of genomics, my story is a little bit of luck, a little bit of nascent insight and, I grew up in a science family, so it wasn’t totally shocking. My father was a scientist. It wasn’t shocking that I gravitated towards a science undergraduate degree. I watched my older brother decide to get an MD and a PhD to become a physician scientist, and that seemed like a good decision. And I, it felt right to me. I really liked the idea. I knew I always wanted to do science and research. I like the idea of combining it with becoming a physician. But up until the point that I graduated with a MD PhD, I was always– I got a PhD in cell biology. I could barely spell DNA when I graduated medical school and graduate school. ’cause I had done my PhD in studying sugar molecules attached to proteins, these things, glycoproteins.

Eric Green: I was very interested how cells try to sort their proteins within to go to different places within the cell to, to kick them out of the cell. And I really hadn’t thought much about molecular biology or DNA science, et cetera, et cetera. then I decided to train in clinical pathology ’cause it was a very light clinical training. ’cause I wanted to get back to the lab as soon as possible. And at the time there was some in my first, my internship year before I then became a postdoc and back in a research lab. There were some early things that were called Molecular Diagnostic Labs where they were just starting to look at DNA rearrangements in cancer by, by, by technologies that we almost laugh at now. But I thought that seemed interesting. And then again, remember when I graduated, the word genomics was coined.

Eric Green: There was a lot of discussion about this thing called the Human Genome Project. And my circumstance, that I stayed at Washington University where I’d gotten my MD/PhD for residency ’cause my wife was a medical student. And just so happened that Washington University had some of the leading founding folks in the field of genomics who were just leaders immediately and really helped catapult the field forward. And so I got to pick a postdoc mentor and I looked around, I said, I haven’t done anything in molecular biology or DNA science at all. And there’s talk about genomics and there’s these great people here. And I went to talk to one of them. And the rest is history. ’cause I just said, wow, this just seems incredible. They were talking about maybe a human genome project starting and they would be really poised to maybe participate. And so I said I should just build up my tool belt.

Eric Green: I knew biochemistry, I knew you know a lot about glyco biology, I knew about cell biology.

Eric Green: I said, oh, I should probably learn a little bit cutting edge work in DNA science or this thing called genomics. So that was, it was a naive idea. And then when I got into the lab, they started a year later talking about applying for becoming a center for the Human Genome project.

Eric Green: It just felt right. My project was right in line with what they needed to do in the Human Genome project. I just said I got all in, I just got completely mesmerized by this. I found myself suddenly working on the Human Genome Project on day one. I was a postdoc. I’m hiring technicians. Washington University was wonderful to me.

Eric Green: They said “You should just run this ’cause you’re, this is great. This is what, you could build your career in it.” And off I went. I had a vague idea if you took me back in time to 1988, when I made the decision what I was gonna do as a postdoc. I had a vague idea that you could connect the idea of having, if we really sequence the human genome, maybe the way we practice diagnostic medicine or clinical pathology, which was my, my clinical specialty, I could imagine some generation, two, three generations from now. Maybe you would have genomics become part of standard diagnostic medicine, but I couldn’t possibly have defined the timeline or the specific steps to get there. And maybe that’s the most gratifying thing of all is that, then I got involved in the Genome project, got recruited to join the NIH at the National Human Genome Research Institute.

Eric Green: Still never, you know, even when you go back 23 years ago, when we finished sequencing the first human genome by the Human Genome Project, it cost a billion dollars. This was still not in the realm of a clinical test that I, as a trained clinical pathologist, knew. But, but you know, we called for the need at that time for reducing the cost of sequencing DNA as a key thing to follow the Human Genome Project. Um, I was the one that helped craft a sentence in a strategic vision that NHGRI out that talked about how we needed to get it down below a thousand dollars. ’cause I knew a thousand dollars was a good cutoff for a clinical test. that sounded great at the time, 23 years ago, I said, wow, I can’t wait till my grandchildren realize that.

Eric Green: I never thought it would happen on my lifetime, my professional lifetime, let alone on my watch as NHGRI director. But all those things happen. So, And even now, I’m just continually surprised, which is how fast things move. Such that, the fact that I am now existing at a, as a CMO at a company that truly is going to help make genomic medicine a reality worldwide is pretty, pretty remarkable.

Eric Green: And I could I can’t take full credit for having the vision, back in 1988 or whenever that, what this was gonna lead to. But I could see the general direction, but I just feel incredibly lucky that it’s all happening in, in my professional lifetime.

Grant Belgard: What other choices changed your trajectory the most?

Eric Green: I think what other choices, if, when you think about it some choices were, mine and some choices weren’t always mine. One of the, the things I always like to tell young people is. By the way, you don’t ever look at more senior people and think they had a life plan where everything was laid out and don’t even think that they even were in control of all their decisions because a lot of things develop and things. Some of it’s just luck and some of it’s personal, right? It could be, what you have to do as part of your family in terms of when you can move, when you could not move. But I never saw myself, I never saw myself going into the government. I just it wasn’t something that occurred to me. And then one set of circumstance, you know, I thought I was gonna be an academician. I thought I’d be at a university forever.

Eric Green: And I I was an assistant professor at Washington University in 1990, starting in 1992. But nobody ever saw it coming that, that Jim Watson, who was the first director of the Genome Institute and launched the Human Genome Project, that he would end up departing NIH unexpectedly, leaving the Human Genome Project without a leader. All of a sudden, Francis Collins gets recruited to take over the Genome Institute and he says, I’m only gonna do it if we can create a government based genomics program called the, it’s called the intramural program. All of a sudden they say, yeah, you should do that. And Francis immediately tried to recruit a bunch of top flight genomics researchers and genetics research.

Eric Green: He called me up and all of a sudden in one phone call, I completely changed the trajectory of my life. I went from academia and jumped into the government having no idea what I was getting into, but it was, it turned out to be a wonderful opportunity. And within two years I was taking on leadership roles.

Eric Green: I took several, I was able to start a sequencing center and then I was even able to become the director of the institute and then things changed for me. A year ago I retired and and I was I retired from federal service ’cause I was retirement eligible after 31 years. even then, I didn’t know exactly what I was gonna do and I, but I was given a new chapter to explore it.

Eric Green: I, my immediate gut said, go back to academia. And I tried to do some of that. I explored some options. I did some consulting. I also started talking to some other organizations for consulting and and I, all of a sudden was approached by Illumina and it’s it really was almost like instantly, oh my God, this just makes complete sense. And and part of it is, I, it wouldn’t have been just that way for any company. I think it was I’m immensely proud of my time of at NHGRI, and I always say it’s like the icon of genomic funders. It is. It is the top flight one. And I feel the same way about, I’ve always felt the same way about Illumina. They are the icon of genomics companies. It’s, to me, the top one. So all of a sudden when they called me and said, chief Medical Officer and the way it was gonna be formulated was just perfect for me and my family. It was a really easy decision.

Eric Green: I just never saw, yeah, I never saw it coming. And if actually, if. And I had been out looking at jobs and really thinking through what do I wanna do? Where do I wanna, how do I wanna do it? And it just, it hadn’t occurred to me. And although I, to, to some extent it, I was very involved in talking to lots of people and it started through a set of conversations I had, including with the person that eventually became my boss here and the CEO.

Eric Green: So I was open to lots of conversations. I just didn’t see it all falling together until all of a sudden it just did.

Grant Belgard: It’s I’m guessing you’ll have to reach far back for this, but when in your career did you first start to think well beyond individual projects, towards bigger systems?

Eric Green: Yeah, I think implicit in your question is just taking a very big picture view. Of and of scientific challenges. And in some ways it, that was part of what I found very appealing about the Genome project. And if I think about my my graduate work, that was very traditional. Although even then most people have one PhD mentor. I had two. I know. I just immediately saw the opportunity to cobble together that and I also would say that even then, I really enjoyed collaborating with people, some graduate students, some PIs.

Eric Green: I just like working themselves, do my project. I was always very collaborative. Then there was, the Genome project was all about being collaborative and all about being selfless and not worrying about, you really were part of this massive enterprise of thousands of scientists around the world. And really wanna point out for, especially people don’t realize that now we just take that for granted that you know now, you look across biomedical research around the world and there’s still some individual labs, lots of them and individual projects, but consortium science is very accepted. Papers with hundreds of authors, very conventional, you know, now we see that all the time, but back then it was almost frowned upon. Or certainly some people just were uncomfortable with it.

Eric Green: I can certainly tell you that even when I went out looking for jobs in 1991, when I became, you know, not all universities were so, you know, they, I literally applied and got interviewed by some universities in 1991 and they were really interested in having somebody who had some genomics expertise.

Eric Green: It was still a very young discipline. And you’d go there and they’d say what are you gonna work on? And I’d say, oh, I’m gonna keep working on the Human Genome Project. And they said, oh, that’s like really cute that you’re working on the Genome project, you’re not gonna build a career on that.

Eric Green: What are you really gonna do when you grow up? And it’s like, no, no, I’m really gonna do this. And it quickly became apparent that was not a university I wanted to go to. I will say over the years, universities have had to decide whether or not they’re comfortable having their investigators be part of these big consortium, because the genome project was the first of many just in genomics.

Eric Green: But now we’ve seen it in neuroscience, we’ve seen it in microbiology. We see it all over. For some science consortium based work is the way to go. And especially as science gets more complicated and more technical, you need a dream team of interdisciplinary experts and you gotta, you, it’s like a Noah’s ark kind of a thing.

Eric Green: You need to have two people that really know genome sequencing. Two people that really know AI, two people that really know biochemistry, two people that really know physiology. This is how you move big things forward. My thinking about what I what, and it’s not for everybody, but it’s, it meets my personality extremely well. I had, I’ve been thrilled and fortunate to be able to be part of consortium as the Genome Project and things like Encode and some other programs like that. But then in 15 years as NHGRI, director and NHGRI is the poster child of running a consortium. We ran a bunch and we even did some that were NIH wide, even beyond our own institute because NHGRI showed tremendous leadership for NIH Common Front efforts like. Human Microbiome Project, like the human heredity in health in Africa, H three Africa program.

Eric Green: that’s, that’s, this is our expertise for many years and still is. NHGRI really is the leader in how to run consortium based science.

Grant Belgard: What did you underestimate about leadership earlier in your career?

Eric Green: What did I under I think I would say what I underestimated about leadership early in my career, which I always chuckle about, is that, you just had no formal training in what you were being asked to do. You know, and whether that leadership is just running your lab or running a small, like a branch or a department or running a center or running, an entire institute is, all of that.

Eric Green: When you really think about even just running your own lab, it’s like running a small business, right? You have to deal with bringing money in. You gotta balance your budget. You gotta, you have personnel issues. You have, those can get complicated. You gotta hire, you gotta fire, you gotta discipline, you gotta encourage, you gotta mentor, as you’re not really taught much of that.

Eric Green: And sometimes you and then as you take on more and more leadership roles, the res roles and responsibilities get bigger and harder. I will tell you, when you’re at the top of an organization like the yuckier the problem, it’s gonna go to the top. So all of a sudden you’re dealing with things you just can’t even imagine you ever had to deal with. You just don’t realize it until you’re there and you real. But on the other hand, what makes great leaders is I think, and I could talk a lot about leadership. But, two things I think that are absolutely essential is one you’ve gotta just be completely ready to be a lifelong learner of skills.

Eric Green: You just can’t assume that you’re going in and you’re gonna just know what you’re doing. You’re gonna learn it because you have not been taught. But the other key aspect of it is you have to learn how to surround yourself by outstanding people. You have to have the ability to just be very comfortable delegating to them, trusting them, and you have to develop incredibly strong relationships with your immediate folks around you.

Eric Green: ’cause it is a team sport. You run something like an NIH Institute, it is a team sport. And when you’ve got a great team, you’ve got a great institute. And when you, and it’s, you gotta cultivate that team. You gotta pick ’em carefully. You gotta nurture ’em, you gotta mentor ’em. And you gotta let ’em tell you when you’re being wrong and you gotta be humble and be ready to learn and adapt. But it’s it’s critically important. You’re only as good as your, as the team that you surround yourself with.

Grant Belgard: Now for the question you know is coming. What about AI? What are your expectations for the impacts AI will have in the coming years and the new and new challenges it may pose as well.

Eric Green: Yeah it it’s just it’s the same question you could almost ask anybody on the street about society. It’s, you’re just seeing it everywhere and it’s remarkable. Here’s what I, here’s what I would say is it’s almost impossible to even claim your an, for me to claim I’m an expert in this other than I, I can just watch what is happening and it is like everything else it’s just changing in a breathtaking way. Practically measured week by week. You can just see this what I would point out is that not that AI is an incredibly important in all the other ways of everyday life that it’s now finding itself. I, you know, I’ve been saying even before this incredible uptick, so I probably was saying this 5, 6, 7 years ago. Is it’s this humbling nature of, of the genome that I spoke about earlier, that it is just so incredibly complicated.

Eric Green: And the more we dig, the more we realize it’s even more complicated than that. And it’s so humbling. It’s only 3 billion letters and it’s just amazing how much it does to create and operate a human being. And then you realize how incredibly complicated non-coding DNA is, for example, and it’s all a language issue that our, the human brain is just struggling getting our heads around that. It just, this is just seems like the perfect marriage. It just seems like the perfect application is to bring this in to be able to see, generate these massive data sets.

Eric Green: So we’re good at generating the data. We’re just not good at extracting the knowledge from it. It just seems like you know that it’s a match made in heaven. But the other aspect of it is, it’s what I’ve seen in particular in the last 10, 15 years, is that it does attract the best minds. And that’s, that gives me a lot of optimism.

Eric Green: So some of the best computational biology, best AI experts really see this opportunity in biology. And then, and that’s just in the genome. And then of course, if you start to pivot us into sort of the genomic medicine, precision medicine realm, you start layering into that other omic data and other data about ourselves that we capture environmental data, physiology data.

Eric Green: And then you marry that with electronic health records. You know, It’s this big bouillabaise of data that, sounds really great. You wanna integrate. Holy cow. The complexity of that data integration process is we’re we actually, it’s not that we’re gonna benefit from AI. We absolutely, desperately need it.

Eric Green: We’re never gonna get through it without it. It’s like, to remind people that 23 years ago when the genome project ended, I remember we tried, at NHGRI, we tried to convene a workshop. we said, Hey, look at us. We’re really cool. We got 3 billion letters of the human genome finally elucidated. We’ve got a big data issue and I remember we tried to convene a workshop and we invited a whole lot of the, some of the best computational biologists from Harvard, Stanford, yeah princeton all, and we were a little disappointed. A lot of people just weren’t interested. They didn’t wanna come to this workshop.

Eric Green: And normally they would come to NIH workshops and when you would talk to them, they’d say, we’re really big data people, this is climate data, financial data, you, 3 billion, you don’t really have that much data. And it was very humbling ’cause we were drowning in these 3 billion letters. That’s completely changed now. And I think you’ve seen it ’cause the technology uptick, lots of genomic data, a lot of other omic data and obviously electronic health. So now the one thing, unlike 23 years ago now, the best and the brightest come and they wanna work in this area because, ’cause we finally got to the big data stage, even though we thought we were there before.

Eric Green: We weren’t until now or until recently.

Grant Belgard: What advice would you give someone early in a genomics or bioinformatics career?

Eric Green: I would probably have a lot of advice to give them the fir, The first thing I would advise them and ’cause I’m, and is: don’t worry. What I first find about young folks these days, they just worry a lot. And I, and I see this in my own kids just, it’s a generational thing. They just worry.

Eric Green: And yet all I see is opportunities. All I see is in incredible, a spectrum of things to be done that, that are really meaningful, important problems that I am quite hopeful there’ll be great employment opportunities, and I know we hit speed bumps and some years are better than others.

Eric Green: And I know we’re in a rough period now but I really do see this incredible set of opportunities and I hope people don’t get so worried. They say, Ugh, I don’t want to go into life science. I don’t want to go into medicine. I don’t wanna go into biology, so forth. So first thing, and then, but the other thing I would say is and part of what they’re worrying about and it gave, was related to the story I told you earlier is worry that you can’t map this out for the next 20 years. Just what I constantly tell people is pick something that really gets you excited, almost to the point where you can’t sleep in the morning. ’cause you just wanna get back to doing what you were doing. And if you’re not that excited about what you’re doing, find something else to do.

Eric Green: And if you’re, if you wanted to be in the, and I think as much data science and AI and all, those are great tools to pick up, but they’re so applicable in so many different areas. Make sure you’re excited about what you’re working on and if it’s, and if you don’t, and also realize the opportunities are gonna change over time and doors are gonna open that you never could have anticipated. And and just be enthusiastic and excited. And if you’re not enthusiastic, excited, find something else. ’cause at the end of the day, there’s plenty of opportunities and chances are if you have enough curiosity that you wanted to get trained in computational biology or data science, AI and so forth, I’m sure you’re gonna find something you’re gonna get excited about.

Grant Belgard: What skills are rising in importance faster than people realize?

Eric Green: You know, well, one of the I think I’d say is I, one thing I get a little worried about is that the, you know, especially around my younger people is the pandemic was very isolating. now there’s a lot of a lot of abilities to do remote work and to be very isolating professionally. And I think one of the skills, a little and I’m allowed to worry a little. I don’t worry a lot, but I, one of the things worry a little bit about, And maybe if in the terms of advice is, don’t get too comfortable just being isolated and feeling like I’m just, I actually think one of the really important things that are underappreciated interacting in person with people. Maybe not every day, maybe not but just recognizing that human interaction, especially in in, in the pursuit of advances in science or in medicine, that human interactions are really important.

Eric Green: Some of them could be fine virtually, but they can’t all be fine virtually. I just wanna make sure that the mixing bowl of strategizing, discussing, debating, interacting, And and developing a relationship I think is so important. All this, one Of the things, and I, again, I say this all the time, I think a lot of people a lot of people who are not scientists think that, um, science is a very asocial profession. They think, oh, they think of the scientist at the bench just pipetting and thinking great thoughts and writing great papers. science is one of the most social things I know.

Eric Green: And I’ve, and my career throughout has been overwhelmingly benefited by interactions and then at the extreme interactions at the consortium level where, you’re interacting with people all around the world and you’re, and I’m just worried about the recent trend of, the pandemic and isolating and the ability to exist and work virtually and interact virtually may take a little of the momentum away from what I think is one of, beneficial parts of science, which is being a highly social endeavor. I just would caution people about thinking the perfect job is one where they never have to interact with people. ’cause I think that the interaction is really important, including in-person interaction,

Grant Belgard: So on, on that note. What do you find makes a mentor relationship most useful.

Eric Green: The mentor mentee relationship. Most useful. I think it’s, I think it’s one Having the mentee’s interest mind and not let anything be cookie cutter. you know what? I always tell people, I and I mentor lots of people in lots of ways. Even today, I’m still mentoring lots of people. I can tell you on my calendar this week, there’s like two or three meetings I’m having mentoring folks the, you know, who are late in their career, middle of their career, early in their career, gets a life being mentored as a lifelong endeavor. I’m still getting mentored.

Eric Green: I, I got news for you when I became. The, The CMO I’m in a whole new world. I’ve been having conversations with some of my lifelong mentors just ’cause I think bouncing ideas off of them. What, it’s not cookie cutter. And the other thing is the most important thing for a mentor to realize is their goal is not to replicate themself in the mentee and make a mini me, to help guide a person who let them be who they are and even if they’re radically different. In fact, what I always tell people when I give ’em advice is have role models and you should have anti role models. and even your role models, there may be half the things that you wanna emulate in your role model, and half the things absolutely not emulate in your role model. You should pick and choose and feel what’s comfortable for you and pursue that.

Eric Green: So I, I just think there’s sort of an open-mindedness that you can’t, and, and I think if a mentor is way too much about I’m trying to make widgets, and this is the parameters of that widget, that’s just the wrong approach to mentoring as far as I’m concerned.

Grant Belgard: What question do you wish people ask more often?

Eric Green: What question do I wish, you know, you mean like when they’re entering their career?

Grant Belgard: Yeah. Yeah.

Eric Green: Yeah I wish, people would ask, or just be show I I actually, maybe I like it when you question the journey about how you got to where you are. As opposed to, because what I sometimes find that’s humorous is they’re, going through the process, they’re so fixated on asking questions like: take a third semester of calculus? Should I take physical chemistry or not take physical chemistry, especially those who are really trying to get into competitive programs, either medical school or PhD. And I just, I just think some of those are sort of the, the wrong question and I just wish they would ask more people about their journeys so that they realize that everybody’s journey is different and that they shouldn’t try to put themself in a trajectory that is well-defined because there’s no such thing as a single route to get to a destination.

Grant Belgard: And to wrap up what are you most excited to learn next?

Eric Green: To be honest with you, maybe the answer I would give is what I am most excited to learn next is how I can be impactful in a whole new setting, one that is focused on the same goals that I had in the previous three decades. And what I mean by that is, I feel so fortunate by having the career that I had at the National Human Genome Research Institute and I, laser focus, if I had to summarize what we were trying to accomplish, especially me as the director, was, can we make genomics truly improve the practice of medicine? I am now in a completely different role in a completely different part of the ecosystem. And yet I’m, we’re trying to do the same exact thing. And, so what I am most excited about and what I really most wanna learn is what can I do to, to help Illumina in their role in the ecosystem. Truly see genomic medicine and precision medicine become a reality.

Grant Belgard: Eric, thank you so much for joining us.

Grant Belgard: It’s been a pleasure.

Eric Green: Absolutely. It’s been a great conversation and good to see you.

Transcript of Episode 83: Jenny Yang

Disclaimer: Transcripts are automated and may contain errors.

​

Grant Belgard: Welcome to the Bioinformatics CRO Podcast. I’m Grant Belgard. Today I’m joined by Dr. Jenny Yang, co-founder and CEO of Outpost Bio. Jenny works at the intersection of biology, machine learning and precision medicine. At Outpost, she and her team are building a platform that pairs automated experimentation with machine learning to better understand how microbial communities shape drug response, nutrition and health.

Grant Belgard: Before starting Outpost, Jenny studied engineering, physics and bioinformatics at UBC and completed her doctorate at Oxford, where her work focused on clinical machine learning. We’ll talk about what she’s building now, the path that brought her there, and what she’s learned about building important things at the edge of science and computation.

Grant Belgard: Jenny, welcome.

Jenny Yang: Thank you Grant, excited to be here.

Grant Belgard: I’m excited to have you. So for listeners just meeting you, what are you spending your time on right now and what are you trying to build?

Jenny Yang: So really we’re trying to make human microbiology computable. So what that really looks like day to day is we have team members in the wet lab and team members on the dry lab side working on machine learning models. So in the wet lab we’re actually, sourcing human derived microbial communities, running perturbation assays and taking multi omic measurements.

Jenny Yang: So think, metagenomics, sometimes transcriptomics, definitely metabolomics. So really trying to understand communities of bacteria before and after. With different chemical compounds and how those microbial communities actually break down these compounds. For example, microbial communities from the gut.

Jenny Yang: How do those break down a drug chemical or a food molecule? And then in the dry lab, we’re building machine learning models on top of this to help create tools that will allow other people to do their own analyses on these microbial communities.

Grant Belgard: What problem keeps you most intellectually stimulated right now?

Jenny Yang: I think I’m definitely motivated by the question of. How can we make meaningful pushes forward in the field of personalized health? I’ve definitely spent a lot of my last 10 years of my adult career looking at human genomics, and I’ve seen incredible advancements in that field. We can identify, exact mutations that lead to certain subtypes of disease and even choose targeted therapies towards that disease. And I really want to understand why despite all this progress, we still can’t push forward improvements in clinical trials and drug development. So the metric I like to show is that 90% of drugs failed 40 years ago and 90% of drugs are failing during clinical trials today. So why have we not pushed forward that metric despite all these huge advancements in the field? That definitely does keep me up. And I feel like the next addition of biological information that needs to be added to this picture is really all the microbes that are in and on our body that affect how we all independently experience health and disease.

Grant Belgard: So for what you’re doing what is the most difficult aspect of the problem from the biology side and what’s the most difficult from the data or a modeling side.

Jenny Yang: Definitely on the biology side it is reproducibility and scalability of these experiments. So when people have been doing microbiome based analyses in their own respective labs, there’s bias that’s going to exist. So it’s been really hard to reproduce findings in one lab to another lab. And then also scaling up these experiments is hard.

Jenny Yang: So if we specifically focus on the gut microbiome, these are anaerobes that we’re dealing with. They have to be in a oxygen free or really oxygen low environment to keep them alive and actually perform experiments. So scaling that process up is very difficult. So when we first started the company, we knew we wanted to focus on both developing a high throughput method that could perform these perturbation experiments, but also make sure we cross validate findings in the lab.

Jenny Yang: So not only do we generate our own data in our lab, but we’re making sure to generate external data sets with other partners just so we can demonstrate that our findings in our lab can be reproduced elsewhere. And in the, on the computational side, we’re dealing with a very high dimensional space.

Jenny Yang: So even compared to human genomics where there’s only about 20,000 genes, the gut microbiome can have up to 150 to 250 times more genetic diversity there. So it’s already a much more. Complex space that we’re dealing with. And again, bringing it back to human genomics between you and I and everyone else in the world, we share 99.9% the exact same human genome.

Jenny Yang: But if we look at our gut microbiomes, we can be up to 90% different. So not only is there huge individual variability, that multiplied with the diversity of bugs that exist in our guts. That is a huge computational challenge. So without the sophisticated machine learning models that we now finally have, as well as the cloud compute infrastructure that we now have, this is just such a huge problem that we couldn’t have tackled before. So that, that level of complexity, I think is what makes it very challenging on the actual computational side.

Grant Belgard: And can you tell us more about the stage you’re at right now with Outpost Bio and what your role actually looks like week to week at this stage?

Jenny Yang: Yeah, absolutely. We have been a company for one year. We raised our pre-seed last July. So that was a 3.5 million round and really the focus of the past year and. And the rest of this year as well was validating our wet lab method and starting to build initial machine learning models and really benchmarking against current state-of-the-art models and benchmarking specific problems. Just to demonstrate the usability and the accuracy of both the data we’re generating and the models we’re, we’re building. My day to day has. Changed significantly from when we first started the company to now. So really when we first started the company, a lot of it was making sure we build up a vision for the company that could inspire people, but also we could see ourselves growing into over the course of the company.

Jenny Yang: And that really is trying to make human microbiology computable so we can improve personalized health. And then it moved into building up our values. So when we hired, we really wanted to make sure that we built up the culture that we were looking for. So not just look for the technical skillset, but also just the cultural fit for the company we envision having.

Jenny Yang: ’cause when you have such a small team, at the beginning, culture is really everything. And I still believe that one year down the road, and now it’s really transitioned into. Just a little bit of everything. We have so many different things going on at the same time. So we have a wet lab in-house that we just opened last week. We have multiple models being trained. A lot of partnership conversations with potential customers, so I have my toes in everything. But the team, is fantastic on, on staying organized, so it’s been very exciting.

Grant Belgard: What decisions do you feel you still need to hold closely, and which ones are you happy to let others own?

Jenny Yang: I very much believe in this quote that I think Steve Jobs and probably a lot of people have said, which is, basically you hire incredible people and you shouldn’t be telling them exactly what to do because you hired them to tell you what the right thing to do is.

Jenny Yang: And I, I really believe that. So we wanted to be very rigorous on the scientific side. So we hired a VP of RD with a lot of experience in microbiology. This is Heidi. She’s trained in incredible labs and she’s just so good at what she does, and I would never want. Want it to be me telling her how to do good science.

Jenny Yang: Like I can point her in the direction that we’d love to go as a company, but I really trust her to be the expert in what the position that we’ve hired her for. Same with Seth and Nathan on the data engineering and machine learning side of things. I’m very happy to let them do the tasks that they think are right to reach the milestones we want to reach. And I think that’s really important because especially at a startup where there’s so many things that need to get done as one person, you just can’t do it all. And you’re gonna have to have trust in your team to, to do the tasks that they’re the best at.

Grant Belgard: What tradeoffs are you navigating between scientific depth and execution speed?

Jenny Yang: I think a big one is the trade off between generating a lot of data really quickly, really broadly, or generating a bit less data, but focusing more in depth on specific verticals. And that is a trade off because we are trying to build a foundation model. And that term in itself suggests that you have a lot of data to build this large machine learning model that’s going to be generalizable. But I do believe because we’re tackling such an important problem that has to do with pushing forward human health and our understanding of it, we have to be scientifically rigorous, especially if the tools we’re creating are meant to be used by other scientists. There’s a level of validation. And scientific rigor that’s really needed for people to trust that you’ve built something that is usable and accurate and if they implemented it into their own work, which is very important that being rigorous and deciding to use our tooling. So especially at this stage of the company where we are trying to make sure we prove that we can do what we say we can do, We are making sure that we focus really rigorously on validating specific verticals, so we won’t go hugely broad and generalizable at the beginning. We’ll demonstrate that we have a method that can lead us there down the road but we need to be more rigorous on proving the science in specific niches first.

Grant Belgard: On that note. How do you think about robustness, bias, and generalizability when the biology itself is so variable and complicated?

Jenny Yang: Yeah, so I think about this a lot. When I did my PhD, I specifically focused on bias mitigation and machine learning generalizability. So from the wet lab side, we are generating data in-house. But we’re making sure to have additional partners generate subsets of data for us. So we’re working with a West coast US institution as well as a CRO, and we’re making sure that they generate buckets of data so we can validate findings externally across sites.

Jenny Yang: And that is really to demonstrate the reproducibility of the wet lab methods we’re using. We’re also. We’re also making sure that we validate machine learning models, so having external data sets from different sites can then demonstrate the generalizability of machine learning models. And we are making sure that we evaluate both metrics of accuracy and metrics of bias because we wouldn’t want to see model performance being dependent on where data is generated.

Jenny Yang: Ideally, it would be based on the actual science itself.

Grant Belgard: What would a truly meaningful win look like for you over the next few years?

Jenny Yang: I think something meaningful would really be demonstrating that the systems we’ve built both in the wet lab in vitro and the in silico systems we’ve built in the dry lab translate to actual h uman outcomes. So another form of validation that we really think about is how can we validate on actual human cohorts or clinical cohorts. And that is something top of mind. So we are making sure we’re working with scientific advisors that are actually running clinical trials or working with groups of patients. We have a project coming up in, in the summer specifically focused on a cohort of patients that we’ll collect samples from, and then make sure that we validate our, both our in vitro and in silico methods again.

Jenny Yang: So I think a huge win would be demonstrating that translatability into the real world. Another win that I think a lot of the team is really excited about is demonstrating that we can actually generalize between, a community of bacteria interacting with a molecule and leading to a certain biological outcome.

Jenny Yang: I think that would be incredible to see that microbiology does follow, some sort of pattern that can be predicted.

Grant Belgard: What turns a biological data set from merely interesting to decision grade?

Jenny Yang: I think that’s a super good question. And it might depend a lot on. Also the question you’re looking at and the industry that you’re trying to implement this in, for example, like there’s a lot more rigorous testing that needs to be done within drug testing versus maybe developing a consumer good. But I think for us it’s making sure that we do focus on certain verticals and then. Validate and calculate metrics that are needed at proof points that are needed for the specific problem, its itself. So there is a little of bespoke kind of tailoring needed for some problems, but I. I feel like it’s in general across all spaces, we would wanna be able to generalize across multiple data sets that are coming from the field, multiple problems that are related to those data sets and demonstrate it with really external partners because we can do a lot of testing ourselves, but when you actually bring it into the real world and it’s in someone else’s hands and not really in your control, that’s where you would wanna be able to demonstrate usability and, a certain level of accuracy there.

Grant Belgard: What does a strong feedback loop between wet lab work and computational work look like in practice in this space.

Jenny Yang: This is super important to us. So when we first started the company, a lot of people told us, make sure you keep your wet lab and dry lab close together and make sure communication is good between the two. ’cause it’s like completely different languages that these scientists would be used to speaking. So for us, we very much believe in having this quote unquote lab in the loop. So the experiments we generate in the wet lab, we make sure people on the computational side understand these experiments are part of the decision making process. Because they need to understand the data that they’re getting and what kind of problems the data can actually solve.

Jenny Yang: Like what kind of outcomes are coming from this wet lab data that they can build a model towards. And then the findings in the wet lab will inform a lot to the to the scientists in the wet lab because. The people from the dry lab will be able to communicate the quality of the data that’s coming out the accuracy of the models where bias exists.

Jenny Yang: So if bias exists in one section, maybe we need to be generating specific data sets targeting that. So we, we definitely believe in just really good communication all across and making sure everyone’s involved in understanding the overall experiment.

Grant Belgard: When you evaluate a possible collaborator or a partner, what signals make you excited?

Jenny Yang: I think people who are really interested in getting to a mechanistic or causal understanding of their food product or their offering. So one example or really any ingredient manufacturer, they wanna understand how, their ingredient or molecule will perform differently across different populations based on their microbial makeup. And also people who are trying to add another level of personalization and provide more of a personalized understanding to their clients. So you can think of people trying to come up with personalized diets for different individuals. Companies that are, that have an app that help you understand your gut microbiome and how that affects what you should eat. So really anybody who’s trying to put a molecule in or on you and wanting to understand the effects.

Grant Belgard: Where do you think the field still confuses correlation with something more actionable?

Jenny Yang: I think there’s definitely a lot of difficulties getting to that causal or mechanistic level. So I think a lot of the field previously has been just focused on correlations. They’ve been saying these communities of bacteria exist and they might lead to better or worse health. They might suggest that you should have these foods versus others. And I think it’s just still very present at that level. We’re really just at the cutting edge of being able to start getting down to the mechanistic level at a scale that just wouldn’t have been, we wouldn’t have been able to do before. So I think there, there is a little bit of an onus on us to demonstrate that this is scalable.

Jenny Yang: ’cause I don’t think people will necessarily be able to implement this high throughput until they’ve seen it demonstrated to a level that they believe they can trust it. So I think there is a lot of work to be done there.

Grant Belgard: What do people building tools for life science often misunderstand about adoption?

Jenny Yang: I think something that I’ve realized over my time in working in machine learning is that. A lot of people immediately expect machine learning models from one place to work really well when it’s brought to an external setting. And I think the attitude with foundation models is one that I believe in a bit more, which is that a foundation model is a starting point to help you fine tune to creating that really high quality model for your own application or your own purpose.

Jenny Yang: I think. Getting to generalizability in such a complex and diverse field, like anything related to biology or microbiology especially. It’s gonna be really difficult to assume that there’s one machine learning model that’s going to work for all. It’s analogous to what we’ve said about the one size fits all model in, in medicine as well.

Jenny Yang: It’s hard to believe that there’s one medicine that’s just gonna work across. Everybody and work every time. I don’t think we should have that belief for machine learning as well. So I think the foundation model space has really opened up the understanding that you start with a model that’s going to bring you to 80% of the way of solving your problem.

Jenny Yang: But if you have your own data sets that are built in the context of your own setting and the problem you’re trying to solve, if you fine tune on that data set, then you’ll be able to build a really strong machine learning model for your own purposes. So I think bringing tools into the life sciences, should not be like just flat out expecting that there’s one tool that’s going to work perfectly across every every team, every problem, every setting.

Jenny Yang: I think there needs to be care into fine tuning these models.

Grant Belgard: Couldn’t agree more. If you could instantly fix one part of the big data ecosystem, what would you change first?

Jenny Yang: I think language modeling, so like chatGPT and all these different large language models that have come out has demonstrated that you can scrape the web of its data and build really strong models that are general purpose and can be used by a lot of people. I think that is it’s a really effective problem and it works for language modeling, but that’s not something that necessarily translates to biology.

Jenny Yang: So I think you see a lot of companies coming out and people building models on all the biological data that they can scrape from the web. And I’ve definitely seen this in the field we’re in right now with the microbiome, but there’s so much bias that exists depending on where you get this data from, that I think it would be very hard to say that you could scrape the web of all this biological data and get to a general purpose model that will understand the problem. Because a lot of these data sets have their own scientific processes. The measuring tools are different. There’s just so much bias that is not controlled that I think it’s a hard attitude to bring into the life sciences. I think you can leverage public data, but I think at the end of the day you need control and confidence over the biological process that you use to generate this data. I very much believe that, especially in machine learning, it is garbage in, garbage out. So as much as you can leverage the public resources, you should, but you should take care into, again, fine tuning, using high quality data that you genuinely are very confident in for the problem you’re solving.

Grant Belgard: What first pulled you toward biology computation and the overlap between them.

Jenny Yang: I’m honestly not too sure. When I started undergrad I was in engineering physics and really working on like robotics and systems engineering. And this was around the time where machine learning was really starting to ramp up and there was a lot of hype around it. And I think naturally I’m just attracted to novelty and learning new things. So I wanted to try machine learning and when I first stepped into it, I was essentially just scraping the web of a genomics database, building a small program that would analyze it, like really simple analysis. And then I just cold emailed 20 different professors at UBC and then. Asking if I could demonstrate what I’ve built. And then Steve Jones he actually just offered me a job after he saw that program and he runs a research lab at the Genome Sciences Center and works very closely with clinicians all working on personalized medicine for oncology. So I spent a summer there and then just fell in love with the intersection of getting to actually work with clinicians to develop something that’s usable by them and would benefit patients but also bioinformatics and machine learning teams on big data challenges. And ever since then I basically asked Steve for a different position every year. I was still in school at this point, so like I would ask to come back in the summer or do a directed studies program with him, and I just kept coming back to the problem and falling more and more in love with it.

Jenny Yang: So that’s just what started me on the journey. It felt like just a positive flywheel that I’ve never jumped off of.

Grant Belgard: Looking back, what early experiences most shaped the way you go about choosing the problems you work on?

Jenny Yang: I, I think I always start with the clinical significance. So when I was at the Genome Sciences Center, the clinicians would always tell us about the problems that they or like the way that tooling could help their processes. There were models that were being built that would try to take genomics information and then predict the subtype of cancer. There was, so that’s like a decision support tool. There were teams that were working on creating annotation systems for automatically annotating histology images and marking out where tumor cells were. And that was part of my Master’s project as well. And I really believe that translational AI can be so effective.

Jenny Yang: You do have to make sure that it works with the clinicians or the people that would actually be using these tools like you. It’s if you come in as just a machine learning researcher, create a tool that you think is useful, it might not actually work with the workflow of a clinician that would be using it.

Jenny Yang: So I think I’m always driven by, let’s start with what can help the people that would be the end user of this tool. And as I’ve moved forward, like through my PhD. And then now with Outpost. I think another thing that really interests me is where are we starting to generate new data sets that we couldn’t have generated before that can unlock insights for more people, more broadly.

Jenny Yang: So I, I definitely have seen kind of the AI for drug discovery and all other biotech fields move towards data-driven analysis. So I know high- quality data that could be used for a lot of different downstream tasks is very valuable because that’s what that’s really where I believe the field is headed. so the microbiome field was really an interest to me at this point because the field of, for example, metabolomics has allowed us to actually study biochemical reactions and get down to a mechanistic level of what’s happening in the microbiome. So that means we can potentially unlock new data sets that we wouldn’t have access to before, at least not at the scale that we could have now which can help solve a lot of problems in a lot of different fields. So I think that’s been a more recent excitement for me.

Grant Belgard: What were some forks in the road that turned out to matter a lot more than they seemed they would at the time?

Jenny Yang: Ooh, this is a great question. I think one decision that we really had to decide on was at what point do we open our own wet So when we first raised our pre-seed funding, we did have the thought that we would just outsource all the experimentation, part of that is cost savings and just efficiency.

Jenny Yang: We assumed that CROs would have the capabilities to just do the experiments we wanted to them to do, and we ended up opening our wet lab t his year. And that’s not something we planned on doing for the first fundraise. And it ended up being really an important decision because what we learned is that what we were trying to do is really specialized.

Jenny Yang: I think we talked to over 60 different CROs and just couldn’t find one that could do what we did. And we had to really work with the CRO that we ended up working with really closely to develop these experiments. And yeah, bringing it into the wet lab, I think it’s really allowed us to speed up. Both data generation ended up becoming a cost saving because we could just move faster. We’re not paying for like external labor of doing the work as well, and we can just test a lot of different variables that we wouldn’t be able to really do with a CRO without adding costs and like extending the time of experiments. So that ended up being a really good decision for us. I think there is something to say about having control over, just what gets done and at what time and what speed.

Grant Belgard: What made you decide that building a company was the right vehicle for the problem you cared about?

Jenny Yang: I had I, so I’ve always been someone who’s been very motivated by the mission that I’m working on, so I’m one of those lucky people where I found a field that I liked really early on in undergrad and have been through that throughout my entire adult career. And I also am someone who had a really good experience during my PhD, so I came outta my PhD still with a positive view of academia.

Jenny Yang: So I was very fortunate for that. I think I would’ve been very happy, doing this type of work in academia, doing it in industry, or doing it in doing it at a startup scale. What I really cared about when it came to making the decision was, how much ability would I have to create vision that I see because I’ve spent a long time building up the skillset and the confidence that I have, I ideas that I can actually follow through on, which meant I really wanted to make sure I was working on a problem where I could have some autonomy over decisions that were being made. wanted to make sure I worked with a team that I really I’m really passionate about what I do, and it’s really made much more enjoyable when you’re around a team that inspires you, that you admire and you trust and are also similarly motivated. And I think if I found, an academic lab that met all those criteria, I would’ve been very happy. And if I, it just happens that I got dragged into the startup world. So definitely a decision that I’m happy I made ’cause I’m having so much fun doing it and I’m working with people that are incredible and on a problem that I think is very worthwhile. But I think one part that maybe I didn’t appreciate earlier on that there are certain constraints within being in an academic lab that I think we don’t have necess, I haven’t felt the same being in the startup world. So for example, like academic labs can really focus on a specific theme. But when you jump into the startup world, I think it really is just what you choose to do as. As a company. Of course there’s other disadvantages, but it’s been a great setting for what we wanna do right now.

Grant Belgard: What parts of your earlier training or work turned out to be unexpectedly useful once you started building?

Jenny Yang: I think one of the biggest, I think one of the biggest advantages of some of my early work that’s helped me now was actually seeing machine learning translated into the real world. So when I worked at the Genome Sciences Center, I could actually see people building the models and then clinicians using them literally the next week. And that was very rewarding. I saw people developing automated tools for annotation, and then clinicians were actually using those tools immediately. So seeing that translation was incredible. Even through my PhD, I worked on a project really closely with clinicians and we actually deployed that AI tool across hospitals and saw clinical validation of that tool in real time.

Jenny Yang: So as people came into the emergency departments, we saw them being triaged using AI and that I think is an incredible experience. ’cause I do think there are a lot of people who build AI models, they publish a paper on that model, and that’s the end of you hearing about that model or seeing the progress of it. And I think that’s a shame because I do feel like are so many ways that AI can help people and support people in their jobs. So by actually having seen multiple examples of that, especially in a field like healthcare and medicine. Gives me like a certain mindset of what kind of rigor needs to go into building AI models that we want to deploy at Outpost.

Grant Belgard: As your responsibilities have changed over time, what did you have to unlearn?

Jenny Yang: That’s a great question. And I think a big thing that I had to unlearn was, I can’t do everything on my own. So I had a very independent

Grant Belgard: PhD

Jenny Yang: where I really got to choose the projects I worked on, drive those projects, and I had to do them from beginning to end. I did all the coding, I did all the research on it.

Jenny Yang: I wrote all the papers, et cetera. I did all of that quite independently, and I of course I had my supervisor and like clinicians and other people on the projects that I could talk to, but I really made all those decisions and something that I’ve had to. Not necessarily unlearned, but like a new skill that I had to learn was delegating, being more organized and combining different people’s outcomes to create the final outcome. It’s very different than being an academia where you’re like much more of an independent researcher. Now we have a team with people working on different aspects, on a really big So That’s been a change.

Grant Belgard: Who or what most shaped your taste in science, leadership, or risk.

Jenny Yang: I think it’s a mix of two people. So Professor Steven Jones at the Genome Sciences Center. I feel like he leads with a lot of trust. He is also someone who lets the work speak. I think he is, someone who doesn’t say a lot necessarily, but everything he does say just is very inspiring, has purpose, has weight and I think that’s a really thoughtful leadership approach.

Jenny Yang: I think it’s also a person, like a personality thing as well, but I really like how he leads with trust. When I was at Oxford doing my PhD, my professor David Clifton, I also think he leads with a lot of trust, which I appreciated. He also was just so efficient. If I needed anything and I emailed him, I would get an email back in a really, like within 10 minutes probably, unless he was in the middle of a meeting. And that just made me realize how important it is. To give people the things they need to do their work really well, especially if you’re in a leadership position. You don’t wanna be the person blocking your teammates from moving quickly or being able to do their job to their best of their abilities.

Jenny Yang: So being present and trying your best organized and answering emails like e even now, like when I see my email fill up. Or my inbox fill up. It can be a little anxiety inducing and I would love to procrastinate it, but I know as a leader especially I gotta just be on top of that because there are gonna be certain things that I just have to answer to keep the machine running smoothly.

Jenny Yang: The other thing about David Clifton that’s really inspired me is that I think he’s he’s a very, he’s a very good listener. And I think Steve is like this too. I think they really enjoyed listening to what other people have to say because if you can understand the people around you, you can make better decisions, especially in a leadership position.

Jenny Yang: I think one of your main jobs is just taking the information you have and making the best decision you can at that time, and that does require more listening than speaking.

Grant Belgard: Was there a point at which your definition of success changed, and if so, what caused the shift?

Jenny Yang: I think my definition of success has not necessarily changed. And what I mean by that is I’ve always been someone who’s just having a growth mindset. I like setting small goals that are achievable along the way, but I feel like there’s always, and celebrate those goals. But there is something exciting to look forward to next.

Jenny Yang: So I think my definition of success is, and when it comes to completing tasks, it really is just finish what you start. Even if you get an outcome that isn’t great, wrap it up. And you can celebrate those small wins. And I think that’s been nice. I’m not of the mindset that you always have something else to look forward to, so you should never feel like you’re finished.

Jenny Yang: I think you should celebrate the small wins along the way, but I do success knowing that you can finish something from start to beginning, even if it’s not necessarily the preferred outcome. You can wrap it up and learn from it and then move forward. And I think now working with a team. My definition of success probably has changed a bit more because now it’s also about keeping team morale up, seeing how motivated other people are, seeing other people’s abilities to do their tasks from beginning to end, and hopefully being part of the support system to get them there. So if anything, the definition of success has just become more of like a we thing than just a independent thing.

Grant Belgard: What advice would you give to someone who wants to build at the interface of biology and computation?

Jenny Yang: I think one piece of advice would be making sure that you bring on the experts in the respective parts, because I really think if you build at the intersection of AI and like biology, healthcare, et cetera, are a lot of important components to it, and. It’s not gonna be able to be tackled by just machine learning engineers or just clinicians or just biologists. I really think it is going to be a mix of all three. And you need to tightly couple the understandings between everyone, like high level enough that people understand the overall goal and how the components connect, but in depth enough that you have the depth of knowledge for each individual component.

Jenny Yang: Because. It’s not a, it’s not a single faceted a field. And every one of those components will be just as important.

Grant Belgard: How should early career people think about depth versus breadth?

Jenny Yang: I do believe in if you’re going to take the leap into building something yourself, I think you should build up a mastery of a, at least a skill set that will contribute to that. So I think when you’re building up that mastery, it’s okay to be very focused on a specific niche but understand where it fits in the general problem a general context of the field. because as one person, you can really only tackle really well, I think, a specific component of something so big, but understanding where it fits can help you, or at least what I’ve found, help you decide maybe what the next step could be or what the next application of what you’ve done is because you’re really building modularly in that sense, rather than maybe immediately having to pivot to a new area. I really wanted everything that I did over the course of my career to build on one another. So it always felt like I was compounding with the skillset I was building and the effect it could have within the greater space.

Grant Belgard: What mistakes do smart people make when they first enter this world?

Jenny Yang: I think entering specifically the AI and biology space and probably other adjacent spaces, but entering it from a startup perspective, I think a lot of people hear that you have to maybe cut corners or take shortcuts and just move fast and break things in order to survive in the startup world.

Jenny Yang: And I think. There is some truth to that. You only have a certain amount of runway and you don’t wanna like waste time in one area and not be able to enough to know the correct direction to go. But especially with biology where experiments matter so much and like a millimeter, a milliliter of incorrect to an experiment will absolutely derail you. I think there is some care that needs to be taken in doing things in an organized fashion at a more steady pace because at the end of the day, that experiment is going to do its own thing. Whether you did it quickly or not. And wanna come out of for example, the first year of company building saying, okay, we have data and that’s all we have.

Jenny Yang: I wanna be able to say we have meaningful data that’s high quality. So I think there’s more caution that needs to be taken coming into this field from a startup perspective than maybe what you typically hear of maybe just a purely software product.

Grant Belgard: What habits or practices have helped you keep learning while the field keeps moving seemingly increasingly faster as time goes on.

Jenny Yang: That’s such a good question, especially we seem to see new AI developments every day, every hour. I think for me it’s, listening, so definitely listening to the people around me. I have incredible AI scientists and microbiologists around me. And my co-founder Alex, our COO, is also just very interested in all the novel advancements coming up, that being able to listen to the people around me really keeps me up to speed.

Jenny Yang: I also like reading. I have news on my phone that I’ll read every morning. I get an update on the news every afternoon, and I just personally. To keep up when I have a little coffee break. The other thing is even though I’m more in a business position now, whereas before I was doing hands-on coding, I’m still very involved with the experiment planning and like reading some of the new articles coming out. I think having some sort of active, component o f my day actually being a little bit more hands-on not to the extent that I was before, keeps me really up to date on what’s happening.

Grant Belgard: How do you think about credibility when the vision is bigger than the proof you can show today?

Jenny Yang: I think it’s very much about how you present the work you’re doing and how you communicate it. So I very much want people to believe our vision, and it’s a huge vision. Like I would love to see a world where can have personalized health outcomes based on our microbial communities in our gut, which is where we’re starting with, on our skin, in our mouth, et cetera.

Jenny Yang: Like I, I think we’re all so unique in that sense that our personalized health recommendations should consider that. And that’s a huge vision saying I’m gonna create some kind of generalizable model to be able to make these recommendations. Huge vision. But I think people believe in that vision.

Jenny Yang: But when we communicate the science we’re doing, we communicate it to the level that we’re achieving it at, which is we’re starting with a focus on the gut microbiome. We’re starting with these classes of drug molecules, these type of food molecules. And we’re doing it, we’re running these type of experiments, so we are going to post open source and we are gonna publish. So we’re gonna be very open with the field to allow other people to evaluate us in peer review. I think credibility is all in how you communicate what you do have. And I think you have to be very responsible there because I would hate for, anyone in the field to put out like, false statements about the truth of the science, especially when it’s in a field that’s so important for human health.

Grant Belgard: What question do you wish more people asked you?

Jenny Yang: Ooh. I think I’m not too sure. I think a lot of what. motivates me are the people that I, I work with and feel like I, I would love to have more opportunities to get to highlight them because especially as a founder of the company I tend to have to naturally be the face of the company and be the person in interviews or answering questions, but I. I think we wouldn’t be where we’re at without the team we have, and everyone is so individual in what they’ve brought to the team that it’s always nice to have more opportunities to, to highlight them because then naturally I can even go more in depth about certain aspects of the company. ’cause I, like I mentioned, being on the intersection of AI is, there’s many components to it.

Grant Belgard: What part of your work still gives you the most curiosity or wonder?

Jenny Yang: I think it is the the translation of real biology, digital form. Because to me I think it’s a very interesting problem to see whether or not we can really digitize or make certain aspects of biology, computable. I think we’ve seen a lot of bodies of evidence that suggest we can. and just really a fascinating problem to me ’cause it feels like you’re taking something that’s this is a poor analogy, but the only one I can think of right now.

Jenny Yang: But you’re taking something that is very 3D and trying to binarize it. And there’s only so many patterns you can represent. Although I do feel like with the rate that. AI is moving maybe this will be really possible, but getting there, I think it is exciting to watch that progress.

Grant Belgard: What belief about the future of biology do you think more people should wrestle with?

Jenny Yang: I think one, I think it really is that. A lot of these biological processes can eventually be automated in the way that we research them or develop data them. I think there is gonna be a very tight interplay between how biological experiments are being done from like the transition between just like humans manually doing it to some automated approach and then looping that in with AI.I think that’s like. It is something that I do believe in. I’m not sure to what extent it’ll be fully automated, but I think people should really wrestle with the idea of which parts can we start automating to make things a lot more efficient. And then where do we still need a human in the loop or how does that kind of evaluation work? And I definitely still, when I think about where Outpost is going in the future, I would love some aspect of automation to really scale up quicker and certain processes to be more efficient. It is something I wrestle with then, where do we make sure we have a human in the loop and what should we automate just yet ’cause we’re not gonna be there?

Grant Belgard: What would you say to your younger self at the very start of this journey?

Jenny Yang: I think one thing I would say to myself is to trust my gut and make decisions, make the best decision you can, even if you don’t have a full understanding of what the outcome will be. I think that’s very important. I think just naturally in a lot of different leadership roles, you don’t have time to get all the information you need, but the quicker you can make a decision and act on it, the quicker you can pivot if you’ve made a mistake.

Jenny Yang: But if the worst thing that could happen is you don’t make a decision at all and there’s no progress being made because it’s better to make the wrong decision, but you’ve learned something and then you can immediately move to the next kind of kind of solution. so I would say. Just move quicker on some decisions even if you don’t have full information.

Grant Belgard: And lastly what do you hope listeners remember from this conversation a week from now?

Jenny Yang: I think it, I think one thing I would love people to really take from this conversation is maybe like an additional peek in their curiosity on how. As humans, we’re not just human DNA, we’re actually these ecosystems, there’s trillions of microbes in and on us. They really affect how we experience health and disease. Yeah, not to creep them out or anything, but maybe just like the curiosity around like how you eat. If you eat something, all these microbes are also eating them. When you use a skin intervention, there’s a bunch of microbes that are dealing with it as well. So maybe it’s just thinking about us more on the systems and ecosystem level than just like a static human being.

Jenny Yang: We’re actually very interesting in that sense.

Grant Belgard: Jenny, thank you so much for joining us today.

Jenny Yang: Thank you so much, Grant. Really appreciate the time.

Transcript of Episode 82: Manuel Corpas

Disclaimer: Transcripts are automated and may contain errors.

Grant Belgard: Welcome to the Bioinformatics CRO podcast. I’m your host, Grant Belgard, and joining me today is Dr. Manuel Corpas, a genomicist, entrepreneur, and educator working at the intersection of genomics, AI, and health data science. Manuel is a senior lecturer at the University of Westminster, founder of Cambridge Precision Medicine and the originator of ClawBio. His work has also included contributions to efforts and tools such as Decipher and BioJS. Manuel, welcome to the podcast.

Manuel Corpas: Thank you, Grant.

Grant Belgard: So for listeners meeting you for the first time, how do you describe what you do right now?

Manuel Corpas: I have been doing bioinformatics since 2000. First as an MSc student at the University of Manchester, then I did a PhD in bioinformatics. Then I moved to the Sanger Institute where I developed one of the leading sort of databases for diagnosis of rare genomic disorders. I did there at Cambridge like four years. Then I started the company, Cambridge Precision Medicine, which was part of the Cambridge University incubator. I did that for a few years. Then COVID came and the wife said, oh, we need to move to London because I want to be closest to my family and so on. Then I moved to London where I got this back to academia as now as a senior lecturer at the University of Westminster in genomics. I lead this program on a new MSc on AI and digital health, which is second year and yeah, with 20 students. So I think that gives you a little bit of an initial set up.

Grant Belgard: What led you to create ClawBio?

Manuel Corpas: So I think the prelude for ClawBio was the bio JavaScript projects, which was an open source community I led around 2014. And the idea was at that time to basically come with a way for stop reinventing the wheel in terms of reusable components that people would need, would want to use for biological data on the web. So fast forward into the internet revolution where we are today, where the idea is that, at least from where I see it, the future looks agentic and knowledge currently in the biological domain, but you could say from many other domains, but I am based on the biomedical domain. So all knowledge tends to be captured by publications, PDFs, which are non native for discovery by agents. And by agents, AI is a software that runs continuously and Alexa, you’ve heard of Alexa, right?

Manuel Corpas: It talks to you and you can ask her questions and it’s able to do specific tasks and it’s all the time on. So the extension to that is that we have now the possibility to use AI powered agents or robots which leverage the power underneath of frontier models, such as ChatGPT, Claude, perplexity, Gemini, you name it. The idea around ClawBio is to come back really to the same situation. So we have bricks of knowledge, which currently are not discoverable. They’re not reusable. They are not reproducible. So how can we make a sort of central repository registry where anyone who is developing this sort of discrete skills can find them in a way that there’s no restriction, MIT license. So anyone can see the code, can push the code. Obviously anything, there’s some guard rails, I look at the code and make sure that it’s audited.

Manuel Corpas: And at the end of the day, ClawBio stands for the idea of using this incredibly adopted new tool called OpenClaw. And I’m going to explain what OpenClaw is. OpenClaw really is the planning between the large language model and then the communication that you have via chatbot, via telegram or WhatsApp or Discord. So the planning in between is done with OpenClaw. And OpenClaw has been gaining a huge amount of traction. So I think it’s like in three months, it’s been the most starred and downloaded in the history of GitHub. And it has, right now, it seems like it’s the sort of base technology on which an agentic AI is going to happen. And so everyone, if you are in the sort of AI domain, you will have thought about agents and agentics all the time.

Manuel Corpas: But if you are not in that sort of domain, the sort of hype around or the sort of heat around agents is that for the first time we have enough sort of capacity via some of the big providers, GPT, Anthropic, and so on, whereby for the first time, they give you the possibility to be able to develop code in a way that basically is hands-free. And that code development, when you transform it into specific tasks and chain them into specific skills, problems, whatsoever, you can then very easily develop a library of repetitive tools, which you can automate without having to be there. So a lot of the sort of vision is how can I automate as much as possible of tasks that normally before we had this sort of elements together. Now we can actually have your agent running them without you having to basically be on the computer all the time.

Manuel Corpas: And you can talk to the agent because it will have already pre-programmed the skills necessary to be able to chat with you. And you can see via your chat and you can ask it questions and it can do and spark code and it can change things just with plain English. So I think that’s really the key of ClawBio. So it’s the application of that kind of paradigm into the biological domain, specifically around reproducibility and also this open source community aspect to be able to share, to be able to reuse.

Grant Belgard: So noting that it’s early March 2026, and this was really just very recently launched, what would you say are currently the most mature skills ClawBio has?

Manuel Corpas: None. I wouldn’t say that there’s any that is mature. The project only has a few weeks and so I cannot claim that. What we do have is some minimal viable products. So some minimal functionality that actually proves the point that you can really do something useful. And I guess for me, the coolest application is a robot, which is currently accessible. It’s called RoboTerri. So RoboTerri’s soul is inspired on my Ph.D. supervisor and Professor Teresa Attwood with whom I did my Ph.D. in Bioinformatics at the University of Manchester. And you can basically talk to RoboTerri, who has already had preloaded my personal genome.

Manuel Corpas: One of the things I tend to demo now is that I can now take a picture of a drug pack that if you could go to the pharmacy and then you say, if you’ve been prescribed this drug, and you would like to know whether that particular drug is compatible with your genetic metabolism for processing the drug. We know that the pharmacogenes influence significantly your ability to respond well or adversely to the normal prescription guidelines. We know that most of the prescriptions and the dosages are based for the average population. And then you have, you know, if it happens that you are not white, northern male European, you’re going to have, and the further you are from that sort of population, you’re going to be more likely to have an adverse drug reaction. Out of those adverse drug reactions are included in your genetics.

Manuel Corpas: So if you are able to just take a picture of your drug, which has your genome preloaded, and then RoboTerri is basically able to identify the active principle in the drug, in the medicine, compare it against FDA approved guidelines for dosages based on your genetics, and then against your current genetic variation, the appropriate dosage based on those guidelines. And then it gives you back a report there in your telephone specifically saying, oh, avoid, or yes, this is good for you. So that’s a lot of minimum viable products I currently have that I demo most.

Grant Belgard: And looking forward to say the end of the year, what would your best guess be about where this project or successors of it go?

Manuel Corpas: Yeah, the reason why I’m now in this place, I’m now in an incubator building. And I’ve been invited by a number of potential investors who are interested in getting further with this project. And so the priorities for me at the moment is that we’re talking not about a sort of normal type of data, we’re talking about very highly sensitive data as it is your health data, your genome data, this is data which is incredibly sensitive. So if you are going to reuse any of those components, you have no idea who the heck have developed them. You have no, the trust component from having a library is very beautiful. Yeah. But when you are now talking about your own personal data or data which is highly sensitive, like patient data, you’re just not going to reuse that.

Manuel Corpas: So what I want to build around now is a little bit like you have seen, for instance, in some of the social media profiles where you have this sort of blue badge where, you know, that person has been certified, that person has, so we could have some kind of certification aspect where if we have that blue badge for any given skill, we can guarantee, obviously for a fee, that specialty skill has been audited against regulatory systems, the European regulatory system and it has compliance 2730001, for instance, that way building that trust for specific skills, which people will pay for, then we can add that extra layer of certainty and trust, which then I think would enable a much more trusted resources. So I think for me, it’s not a question about how big this is, which we’ve already had at least 40 contributions, and we can see that the project is really being very popular.

Manuel Corpas: But I think the issue now is how do we make this into something that can be trusted? That’s really where my current thinking is.

Grant Belgard: What do you find people most frequently misunderstand about AI within the bioinformatics community?

Manuel Corpas: That they have no idea of how quickly things are changing and the power of new versions from 5.2 to 5.3 people think there’s a small incremental change. No, this is 10 times better. People don’t understand two things, the exponential growth in terms of capability that we’re experiencing every three months. And I don’t know about you, but I work completely different as I was working three months ago. Yeah, if you are like me, we’re running like tech agents at the same time, any given time and all these instances. I don’t look at my emails first. Now I look at what was the way I left it with that agent last night and see where we are at. So that’s one thing. And so I gave a presentation at the London Bioinformatics Meetup on the 26th of February. Yeah, that’s nearly a month ago. And I raised it. It’s meant to be a community of bioinformatics practitioners, right?

Manuel Corpas: You have people from major institutions that you know their names. I’m not going to say any names. I asked them, does anybody know what Claude code is? No one. And I was like, oh my God. And this is meant to be it. The top people in my field, I think that, or maybe I’m just mad. That’s why I said I was mad because the relentlessness, the pace, the acceleration of at least how I live on a daily basis, how things are changing, it’s just absolutely, I don’t even sleep because I have this sort of situation where I see this coming. And to be honest, I made a very conscious decision. I have two options. I either stay, unless somebody else makes the decisions for me, as it is happening right now, and I’ll give you some examples in a minute, or I actually invent something and I shape somehow in my own very little small niche, what’s going to be the future.

Manuel Corpas: Because otherwise, unless I invent it, I’m going to be out, as simple as that. And for me being out, that to me is, I can’t live with that though. The example. So I was doing a benchmark of the main frontier models. Yeah. Gemini 3.0, Claude Opus 4.6, Sonnet 4.6, GPT 5.3, Deep Seek 3.1. I think it was a 3.0. And I was basically benchmarking for each of the main World Health Organization diseases that give greatest value to humanity, how well each of these models are able to query back some sample research, output research I had taken from biomedical data is called PubMed. Basically, PubMed has all of the biomedical literature index and you can query PubMed for Ebola, let’s say, or for type 2 diabetes or for ischemic disease or whatever.

Manuel Corpas: I was like, I was actually looking at the 170C global burden of disease, World Health Organization diseases, which include Zika virus, Ebola, I don’t know. So Claude wouldn’t let me query them for those diseases. I know that they’re doing it in the right way. Okay. So these are guardrails. But I’m like, who are some anonymous software developers, Silicon Valley, to decide on what I think I should be doing? Because I’m sorry to say, I’m a bona fide researcher and you cannot or should not dictate my freedom. I know this is a very silly example, but I think this is just a taste of what could happen very soon. We’re not talking about two years. I’m talking about months down the line because one month is a year in AI speed.

Grant Belgard: So in this very rapidly changing environment, how do you decide when a total workflow is trustworthy? What do you look for?

Manuel Corpas: So first thing is transparency. Can I, can I actually look at the code or can I, is it being, are there any black boxes basically? Yeah. So if there are black boxes, then I don’t trust it because there’s no transparency. And unfortunately, most of the LLMs I use, they are black boxes. So how can I trust something which is a black box? Transparency is one. Secondly, I want to see the faces of the people. I don’t trust a project where it’s just an anonymous community. I want to see the people behind it. I want to see who are the developing, who is making the decisions. Because if I have a problem, I don’t want to have your LLM power chatbot to give me one of your recommended arbitrary sycophantic answers. Third, I want to see, is this scientifically grounded? Is it, what’s the thoroughness criteria? I want to see the track record of the people. I want to see their LinkedIn profiles.

Manuel Corpas: And I guess the irony of this is that suddenly the human side of their project becomes so much more important. And I was having this conversation earlier with one person here at this, and suddenly trust becomes paramount for anything. It’s all about trust now, but it’s not about technical prowess. For me, non-negotiable asset, which is cause, is trust.

Grant Belgard: How do you think that will evolve over the next few years? Because the kind of technical side of things is, yeah, I feel like the kind of presentable on this is massive for anyone paying attention.

Manuel Corpas: To be honest, I have a lot of difficulty even predicting, understanding the present, let alone the future. I think that for me, yeah, because I don’t know about you, but I think we are all drowning and that we’re all unable, humanly impossible to keep up with everything that is happening. I, so I can’t predict technology, but what I can see that what would be the tools that will help me be better prepared for the future is something that I live on a daily basis. And to be honest, these are things which I had not paid too much attention until now, but the principles that I know will help me are the following. One, don’t get overwhelmed. If you can just do 1% of catching up every day, you are doing fine. I’ve seen it because in three months time, that is not 1%. It compounds.

Manuel Corpas: And I have seen that just by every day listening to podcasts like this one, where you have trusted people, whoever you feel trusted, yeah, you can trust that. And you keep your own environment because right now it’s becoming less and less important on where you are at, as long as you have access to the right information channels and you keep connecting on a daily basis, that’s good. The second thing is that I have, I’m training this sort of gut feeling that my personal intuition. Whenever I get a result now from an agent, from an LLM, it’s not, I don’t even read it. I don’t even necessarily, it’s more to do with the gut feeling. So I am now having a lot, making decisions a lot more with my gut than my brain. Because I feel that the logical part is sorted now and it’s more to do with, not so much with the logic of intelligence, but how it feels.

Manuel Corpas: And then the other aspect that I have now accepted, and I think it gives me a tremendous competitive advantage in terms of the principles for the future, for how things will evolve, is I feel unstoppable in the sense that for me, the intelligence is now, is not the limit anymore. It’s my capacity to ask questions. For the first time ever, I can ask, and this is already cliche, Peter Steinberger says that the creator of global, you have an infinitely patient machine that will be able to explain things at your own level. There’s no excuse now not to want to learn. And also I find another thing that is becoming cliche as well is you have this situation where some people say, oh, this is going to atrophy your brain because it’s making all these decisions for you and they’re thinking, well, that’s a valid option.

Manuel Corpas: But I can tell you that you can flick the side and harness it to even encompass problems that you would have never thought you would be able to develop before. Because now if there’s something you understand, you can always ask. And so that’s for me, my strategy, I cannot predict the future. I cannot understand the present really, but I have this sort of inner guide compass, which is now more important than ever, that I think keeps me, to be honest, ahead of the curve. And at the end of the day, the adoption curve is here. So as long as I’m ahead of the adoption curve, I’m going to be fine.

Grant Belgard: And how do you think about maintaining scientific rigor in this world? It seems like it takes more time to consume and digest outputs than it does to create them.

Manuel Corpas: Yeah. So obviously we’re being swamped. So my approach is very simple. Have you heard of the dragonfly method? Right. So dragonflies are the most terrible predators you could ever think of. They were human size. We would all be dead. They wouldn’t be humans, right? They are so precise. Why? Because they have 10,000 lenses. In other words, they have 10,000 different perspectives from which they can really calibrate their environment. The way to survive this sort of lack of validation or verifiability of having your work, you can train your LLM to judge it as a reviewer. Then you can have the perspective. You have another lens that comes, which is the founder, another lens that is the patient, another lens, which is the general party.

Manuel Corpas: And so I’m going to give you complimentary perspectives that allow you to, for the first time have really, if you trust your model, I guess that for me, the trust comes from what model do you trust? So if you have a model that you trust, and the only way to trust that model is by using it and by tinkering, by testing it, by seeing, oh, this one is better for emotional empathy. Oh, this one is better for generating figures. This one is better for integration with this particular tool. And that’s the kind of knowledge that you can only guess if you are invested into embracing all these different technologies. And you need to stop being a user. And now if you really want to ride the wave, you must become a builder. That mentality has to change and that will give you the necessary confidence to be able to understand what the shortcomings are. What’s the best way to prompt?

Manuel Corpas: Because as I said, every tool is going to have their own quirks and not even the providers of these tools understand it, they don’t even understand it. They can only have a, some kind of general frameworks and they have their guard rails, that there’s a huge space in between that remains unexplored. And that is the right strategies because all of these models are all going to be generalists. But if you have that specific, very specific expertise that you had in my case, bioinformatics, let’s say, that I’m going to win in bioinformatics for these very specific genomics tasks, whatever. I’m not going to win at the many other general things that are out there. And that’s where I see for myself anyway, the opportunity.

Grant Belgard: How can others get involved in ClawBio?

Manual Corpas: So just go ClawBio.ai and then it will have pointers to the repository where the code is and then just tell your favorite LLM or tool to help you understand what the code is. And so you have your own personal tutor and that will be the way. The other thing is that obviously we’re now developing channels, we’re doing hackathons here next week in London, we’ll develop ways for people to become community and through email lists, through Discord, through, as I said, we have not even had time to have all of these different options, but I think that the best way is simply to just go to the website and tell your favorite chat book to walk you through it and then start question. Break it. If you break it, I’ll give you, I’ll buy you a gift. This is meant to be, stop thinking about messing up, please messing up. That’s the only way you’re going to learn.

Grant Belgard: So changing track here to talk a bit about you and your own career. Was there a moment when your work stopped feeling like a series of separate projects and started feeling more like a mission?

Manuel Corpas: It’s always felt like a mission since I had very clear, I wanted to become a scientist and the current projects are simply an expression of this relentless need to somehow express the all of this energy and all of this feeling I have for the world. I think it becomes like really like my own temperament. I must say that I’m a little bit in the autism spectrum, so I get quite obsessed with things. And right now I’m obsessed with AI, I’ve been obsessed for several years now on AI and now I’m obsessed with global and really harnessing the power of agentic AI. Because as I said, I see a lot of potential, I see a lot of danger, and I see a lot of people who very in a very, in a good way, they’re trying to impose their values, which I know I may not necessarily agree. And some people talk about super intelligence as the end goal.

Manuel Corpas: I don’t think that’s necessarily one that really excites me because super intelligence for me, it sounds a little bit like you exclude some people. What if these people are disabled? Some other people are talking about super abundance as the end goal for this revolution. I don’t agree with that. That sounds selfish and materialistic. For me, what really the purest purpose for me to pursue this obsession is what I call super enlightenment. If we take into the history, into account the industrial revolution, which is the closest thing I can think of to what’s happening today. You had these people like Nikola Tesla, Immanuel Kant, or even slightly earlier, Isaac Newton. So these were people who saw advancements as a way for growth, not in terms of wealth or power, but as wisdom about really a better understanding of yourself and the world around you. And I know that there are real dangers.

Manuel Corpas: And then this is the fact that you could have super intelligence ruling the world. I understand that there have to be people thinking about those problems. For me, frankly, it’s a bit of a distraction. If I focus on the negative, I prefer to focus on the things that are meaningful to me. It’s more of an integration of the technology, even with a spirit, even if I can say that. And it’s weird because you have these new agents that are so [?]. It’s becoming somehow spiritual, even if we don’t mean to do that. Not that I’m necessarily a religious guy, but I think that this meaning aspect, this need for me to be authentic, value oriented, which sounds obvious. I think it’s now ever more important than before.

Grant Belgard: Looking back, what risk are you glad you took?

Manuel Corpas: I don’t think I have risked enough. I wish I had. I’m a risk taker, but I don’t think it’s enough. I guess for me, I had this sort of pathological state of mind when I’m never satisfied. So I have a constant sense of dissatisfaction, and that sort of drives me to want to improve, I guess, combined with my obsessive behavior for things I feel really passionate about. I am not taking more risks because of people I love, whom I care, and I would be taking more risks if I was on my own, but I think I would probably be dead by now.

Grant Belgard: What skills matter most now that AI is changing how technical work gets done?

Manuel Corpas: I think some people say curiosity and inquisitiveness and asking questions. I’m going to go one level beyond that. And I actually want to say it’s about your attitude, and it’s about your own grounding and having a very, very clear sense of compass in terms of what matters to you and why. So your internal compass, why should you be doing what you are doing? Finding the right motivation drive that gives you that inner fire that keeps you going is for me the most important thing, not just from now. This is just a new reincarnation, but I think that that is something which will never change regardless of what technology we’re surrounded by.

Grant Belgard: What makes someone genuinely strong at interdisciplinary work?

Manuel Corpas: Really, not being afraid of showing your vulnerabilities and being prepared to swallow your ego again and again by people who know more than you. And I think it’s also an essential appetite for learning.

Grant Belgard: Speaking of learning, there are lots of ways to learn. What specific habits in that space have compounded the most for you over time?

Manuel Corpas: Studying. So for me, a day that I don’t spend, I don’t mean studying on the computer. I need my physical book, smelling, it may sound a little bit old school, but if I am on the computer, I get distracted a lot. And so having that sort of discrete physical medium I can touch, there’s this connection, a physical connection, a spiritual connection with that work. And so for me, a constant habit of, one of my vices is I keep buying books from Amazon, which I never read, but it’s one of the things that gives me most pleasure, just buying books. I know that I’m going to read maybe 50% of them, but I have them everywhere. I have one book in my car. If it happens that I’m in the underground and then I don’t have anything to read, I have another book in my, in this wallet, I have a book in that. So I always have a book around me, which will bring me back to my origins.

Manuel Corpas: And the origins is scholarship at the end of the day. And that scholarship for me is not given by the computer. It’s given by that sort of quiet, calm, no noise place. In this case, in the early morning, 6 AM, 5 AM, when everyone is sleeping, everything is quiet. I have my little lights, my pile of beloved books there, and just enjoy that sort of moment of solace and connection with learnings, new, old, ancient, which to be honest is for me, was the essence of civilization or our civilization. That’s why I don’t think that AI necessarily is going to make us less intelligent or less able. Obviously, some skills like having a calculator, now, if I want to have a sort of complex sum, I use a calculator. Yeah. Okay. But that doesn’t mean that I’m going to stop being less intelligent because there are other things, as I said, this gut feeling, which I’m constantly training.

Manuel Corpas: So you are that, you harness that technology and internalize as a new Swiss knife artifacts, which now becomes part of him.

Grant Belgard: And lastly, what advice would you give your younger self?

Manuel Corpas: Don’t doubt yourself so much.

Grant Belgard: So Manuel, this has been a great conversation.

Manuel Corpas: Yeah, absolutely.

Grant Belgard: Thanks.

Manuel Corpas: Thank you.

Transcript of Episode 81: John Connolly

Disclaimer: Transcripts are automated and may contain errors.

Grant Belgard: Welcome back to the Bioinformatics CRO podcast. I’m Grant Belgard. Today, I’m joined by John Connolly, Chief Scientific Officer at the Parker Institute for Cancer Immunotherapy, also known as PICI. PICI’s mission focuses on accelerating breakthrough immune therapies by bringing researchers, tools, and collaboration structures together to move faster from scientific ideas to patient impact. In this conversation, we’ll cover what John’s working on right now, how his career path shaped the way he leads, and practical advice for anyone building in translational science, especially where data, bioinformatics, and real-world complexity collide. Let’s jump in. So, John, for listeners meeting you for the first time, how would you introduce yourself and what you focus on these days?

John Connolly: Yeah, first of all, I just want to say thanks, Grant. It’s a real thrill to be on the podcast with you and to catch up. John Connolly, as you heard, CSO of the Parker Institute for Cancer Immunotherapy. I guess the way I’d introduce the Parker Institute is, first of all, by saying it’s a 501C3 nonprofit. We’re a cancer charity. So we, as you heard, our mission is really to make all cancers curable disease. And the organization really lives to that mission, very much focused on funding fundamental research in the field of cancer immunotherapy, and then really creating what is the highest concentration of immuno-oncology expertise in the world, pulling all of these experts together, betting on these people, not on projects, but on the people and what they do, and creating these 14 centers around the US. They’re full centers or what is known as EMR centers.

John Connolly: And this really gives a lot of autonomy to those sites to do that fundamental research. But in describing the Parker Institute, it’s not just a grant agency that gives money out. What we’ve done is we’ve really knitted these organizations together, sort of a network of networks together above each of the individual institutes. What we’re trying to do is break down those barriers, break the silos, so you’ll be able to exchange data, information, projects, people between the different institutions. So we pre-negotiate confidentiality agreements, material transfer agreements, everything is there. So you can just stand up in the middle of a presentation and send a vector to the next lab just immediately, or open up and really have these substantive discussions about what’s next in the cancer space.

John Connolly: And that’s really a major part of the institute is that funding, is that research, is finding those young investigators, building out those centers around the US. And then the next part really is translating that into real clinical trials in the real world space. And so to do that, I think there’s no better place for blue sky ideation than academia. I mean, this is kind of what it does well. But one thing it doesn’t do well is prioritize programs and execute in the clinic. So what we’ve done is we basically created a venture philanthropy arm that spins out biotech companies based on the technology that’s coming out of the network. So getting best in class from Stanford, MD Anderson and Dana Farber and pulling them all into one company and then putting that under project management. So we’ll fund those companies and actually staff them.

John Connolly: We’ve got a new co-build team that can jump in in operational roles and then push that technology out and really test it in the clinic under project management. So you actually see it executed there. And we have a business development arm that can then work with pharma to get this out to a much broader community. And so with that, I think you really do have to have that. If you want to accomplish such a ambitious goal like all cancers curable disease, you really need to have a line of sight from the blue sky kind of science all the way to commercialization. And when I say line of sight, you need to have operational efficiencies in each of those areas. And that’s really what the Parker Institute does. It’s been around now for 10 years, 2016 we started it. Really everything I just described was the idea of Sean Parker.

John Connolly: So Sean really saw this vision of there’s, these silos are really preventing the immunotherapy from really reaching its full potential. And there has to be a way to break down these silos so that people, the best people at MD Anderson can be working shoulder to shoulder with the best people at Cornell or Memorial Sloan Kettering.

Grant Belgard: What’s the problem you find yourself thinking about repeatedly right now?

John Connolly: One of the biggest problems is we’re in the middle of kind of a strategy review for the next 10 years. And so it’s thrilling to really talk about that. I’m working with Ira Mellman and who’s recently joined the Institute from Genentech and to just think about what that 10 year strategy really looks like. And it’s funny to be on this podcast, what I’m thinking about is how we recruit more kind of AI native people into the immunotherapy space. And I think the way the approach that I think works and that we’re taking is not to train people to become kind of proficient to think in this AI kind of first integration, but really grab people that are in the tech space here in San Francisco or in Asia or other where and then turn them into immunotherapists. I think that’s for me the biggest thing.

John Connolly: So we’re excited, but it’s something that’s because of the way we’re thinking about strategy, we’re really pushing forward. So that’s one big one that we’re thinking about.

Grant Belgard: What’s something about cancer immunotherapy that’s widely misunderstood outside the field?

John Connolly: Yeah, I think one of the things for cancer immunotherapy is this idea that, so there’s a lot of expectation with it. I think the, and it’s that you just simply sort of induce an immune response against the cancer and the cancer would suddenly go away. It’s a difficult thing to do. As much as we’ve created incredible successes in the field, and I really mean that, I think the advent of checkpoint blockade is arguably the biggest advance in cancer in 5,000 years since the advent of surgery. I mean, like really. So it works broadly across multiple tumors. It’s curative in late-stage metastatic disease. There’s just nothing else like that. This chemo doesn’t do that, nothing else does that. So it’s phenomenal. And when it works, it works spectacularly. The downside really is that checkpoint blockade works in 20 to 30% of people, and that’s it. In certain cancers, it works very well.

John Connolly: In others, it doesn’t. And so getting a better understanding of why that is, being able to predict, requires a deeper understanding of how the immune system really works. And so I think that’s a big one. The other is this idea that it’s a free ride, right? There’s any kind of novel therapy that comes in that overactivates your immune system. There are obviously side effects, and mitigating those side effects and amplifying the effective nature of the tumor, anti-tumor effect, is totally essential. And that’s another obviously big area of research.

Grant Belgard: What’s a recent moment where you thought, this is why we do the work?

John Connolly: Oh man, this is, luckily for me, this happens a lot. But we’ve recently taken, a lot of efforts gone into immunotherapy, tackling big cancers, man. Lung cancer, all over the place, right? So we’ve got tons of programs and projects in that. Big efforts in breast cancer and others, and they’ve made true progress, like phenomenal progress. But what I’ve recently done is taken a step back and challenged the network to live to the mission, what I said was all cancers curable disease. Not all cancers, just treatable disease, okay? So, I mean, let’s see what we can cure. And so actually doing a scoping, looking at cancers that are highly responsive to immunotherapy, right? And saying, let’s move beyond just treating those cancers. Like rare subtypes of melanoma with Antoni Ribas at UCLA and Paul Nghiem doing Merkel cell carcinoma up at the Hutch.

John Connolly: Actually saying, guys, what if we actually took all of these incredible tools that have been developed for these very common cancers, like lung disease, like lung cancer or breast cancer, and instead focused on rare cancers that we know are gonna be responsive, and actually cure these things. Let’s tick some boxes here, let’s get these things off the list. And so phenomenal work, again, from Tony Ribas, recent publication in rare subtypes of melanoma showing just cures. It’s really, really spectacular. And I’m proud to say we’re putting a clinical trial together at Dana Farber now, which is a combination cancer vaccine for Merkel cell carcinoma. We really dove deep into, can we induce a really potent immune response against this thing that’s highly responsive to checkpoint blockade?

John Connolly: Can we then clone the T cell receptors, ship those to the Hutch and have cells ready for those patients so that if they do recur, they can be treated immediately and cured. So I think just eliminating cancers, like truly curing a whole class of cancer is something that is really near and dear to my heart. I think it’s just on the horizon.

Grant Belgard: How do you translate such a big mission into a concrete research strategy?

John Connolly: Well, I mean, the first thing is some humility, right? So saying that there’s so many things we just don’t know yet, and we don’t know about the immune system, about how cancer interacts with the immune system, and going to find those people that are really focused on those questions. I think in building that networked based foundation, I think this is an essential portion. It’s really the value of the Parker Institute. It’s just the people and the incredible investigators. And then asking them, guess what is it about the current funding mechanisms or the current mechanisms within your university that are holding you back? You talk about all these projects, but it’s been 10 years and we still haven’t seen something come out. And a lot of this has to do with grant funding, review process, bureaucracy, the slow nature of all of this.

John Connolly: And you and I both know that just the way the concentration of expertise works is these centers attract good people that attract good young people to kind of get there. And so what we’ll do is we’ll come in and bet on those people and say, look, your area of focus, let’s give you a block gift, largely discretionary funding, right? With some obviously project guidance through a steering committee. But that’s it, it’s really up to you. How do you want to spend this money? What do you want to do? And then do it in perpetuity. So they know they’re never going to have to worry about that money again, because it’s coming every three to five years, there’s another gift coming, another gift coming. So that takes that constant grant writing off the table for these academics. The second is when we do have steering committees, at the end of the committee, the money immediately goes out.

John Connolly: So it’s just a review, quick review, and then yes, no, modify it, yes or no. And then the money’s out the door at the end of the meeting immediately. So for all those funded projects. And so that kind of speed, it just changes things. I can just speak from experience, if you’re a PI in a lab, and you want to have the ability to, you’re thinking all the time about what your areas of focus constantly, 24 hours a day. And you want to be able to go out on a bike ride and come up with an idea and immediately hire two postdocs to work on it. Like this is what the institute’s kind of funding brings. And it’s just an essential kind of thing that just changes the dynamic.

Grant Belgard: How do you decide what not to do, even if it’s scientifically exciting?

John Connolly: So, I mean, the focus on immunotherapy, even if immunotherapy kind of is turning out of favor in like the investor community or the pharma community, I think this is staying on mission. This is really extremely helpful. So no matter whether it’s a project that we’re funding, or if we’re building a company, or we’re looking at an investment opportunity, or doing a partnership with pharma, the first thing I always ask is, is this mission? That’s all, is this mission? And if it is, then we’re aligned. So we can work together as companies, we can move forward, we can invest in that company or build that company, we can invest in that project or that person. And so I think that that really is a North Star for us, is this idea of all cancers curable disease and a focus on using the immune system to get there.

John Connolly: The other is, there’s a certain, staying in touch with the network, doing regular site visits, getting out there, is really essential to understand where the momentum in the field is going, and where the early results from early stage clinical trials, the innovation in this space is going. So that also kind of helps an idea of which direction and funding is going to go. The other is, I think some of the best ideas out there are coming outside our space. I mentioned AI and other things, but there’s those inventions that come in sideways to try to solve problems that people have been banging their head on for a long time. I look for those all the time. I also look for contrarian troublemakers too. I love those people that’ll jump in there and the worst thing you can have in a network is just everyone saying the same thing.

John Connolly: You definitely want people that’ll go against the grain and shake things up.

Grant Belgard: What are the biggest friction points that slow progress in multi-team science and how do you try to reduce them?

John Connolly: That’s a great question. So I think the biggest kind of hurdles that you have to overcome, and this is probably true for any organization of our size, is that bureaucracy that I talked about. In worst case scenario thinking, this idea that you have to constantly worry that you’re working with another investigator and they’re going to take your ideas or the IP or something like that, or that that university is going to have a better position than a new company build. That kind of worst case thinking just sort of breeds mistrust and what it does is it eliminates opportunities because you end up just taking a defensive position. You see this across a lot of organizations. This is why it’s really essential.

John Connolly: When you understand the Parker Institute, you understand that it’s a network of people, people that trust each other, the people that get together twice a year in these really intense retreats that we do, sharing unpublished data and just getting it out there and really trying to show the best in class. So that trust within the network, it’s kind of a secret sauce. It’s almost impossible to kind of recreate just with money. It’s something that these guys see each other as Parker Institute investigators before they do Harvard professors. And I think that that’s just an important thing that came with the great work of Jeff Bluestone when he built the network, as the first CEO of the Parker Institute, Sean’s vision of building this and pushing it out. And I think the work that my team and I have done pushing it forward.

Grant Belgard: Yeah, how do you design incentives so that collaboration is real and not just a slogan?

John Connolly: This is, so it’s really matching, it’s kind of impedance matching on capability. I think that’s really what it is. Because I don’t think that collaboration all the time is such a great idea. We fund projects that are very similar to each other in individual labs, like real competitive projects that I’m happy to fund if we think we need to double and triple down on the idea. One lab’s not gonna solve this problem. So even if there are very similar ideas and things like AI based protein design or TCRT therapy, there are programs that are extremely competitive labs. These guys have been competing for decades and I think the Parker Institute isn’t gonna solve it. But I think this impedance matching concept that I was getting to is really the real deal. So if you have someone like Chris Garcia at Stanford, Chris is a genius and he was successful long before the Parker Institute.

John Connolly: But one thing that the Parker Institute brings is, now what Chris is doing is he’s building these incredible proteins, these incredible systems, but then they’re immediately getting transferred over to Carl June in Penn, who’s then taking them into the clinic and moving those forward and really seeing Chris’s work applied to medicine. And I think that that kind of matching of capability where you’ve got the people that just live in structural biology that then pass it off to someone who’s a development expert who then passes it off to someone who’s a clinical execution and do so seamlessly because it’s exciting. This is how those projects really move forward. And again, touch base, just always going back to that touchstone of mission, is this something we wanna work on?

Grant Belgard: How do you measure if your strategy’s working, what metrics matter and what metrics can be misleading?

John Connolly: Sure, for sure. So one metric that can be misleading, I think, is publications, so nothing against it, but if you sort of get the absolute top people in the field, you’re gonna get a lot of publications. I guess just kind of, because that’s a metric of universities. And it’s a good one for universities. So it’s very important to get out there to communicate, to spread the word and to excite and attract other people. But for us, I think that some of the key metrics are this sort of full circle. So I mentioned the idea of moving blue sky science and pushing things forward and paradigm changing kind of ideas into execution in the clinic and then into distribution through pharmaceutical companies or through large biotechs to the greater patient community.

John Connolly: So for us, the first thing really has to go back and ask how many patients are actually benefiting from the research that’s happening at the Parker Institute. For me, that’s a big one. And that’s just really looking back on, and we have hundreds of clinical trials with ideas and therapies that have come out of fundamental research at the institute. We keep really good track of that and just see how things are going. I think that’s, for me, that’s probably the biggest one. And I think this is, it takes a lot of time, unfortunately, to get something that someone that’s City of Hope or UCLA comes up with and then translate that into a phase one clinical trial, execute that into the clinic, and then really begin to push it forward, benchmark it in a phase two or three trial and actually see that applied to patients. But this is really what it’s all about.

John Connolly: There’s wonderful organizations who are here to fund just sort of, just fundamental research, right? And the NIH is clearly the global leader and the actual backbone of all research funding and progress that we’re making in the field. And that’s an important and incredibly amazing thing, but that’s not what the Parker Institute is doing, right? Mission is all cancers curable disease and we do what it takes to get there. And so for us, when we go back and check, we ask that question, how well are we doing by looking at what kind of impact in the patient population we’re making?

Grant Belgard: What are the failure modes of cross-sector collaborations and what guardrails help?

John Connolly: Yeah, I think, I mean, some of the failure modes, I mentioned that trust idea, right? So anything that threatens that, because when it’s working, when people are openly sharing data with an expectation and understanding that working together is better than working in silos, that anything that kind of comes in and threatens that is problematic. I mentioned that I love taking contrarian troublemakers, but I don’t want a room full of them. So it’s like, for me, that’s really the big take home is making sure you maintain that network effect. And to do that, you have to listen to the network too.

John Connolly: So one of the most important things about the job at CSO is getting out to the sites, meeting regularly with the center directors, meeting regularly with the young investigators that are coming up, and talking to them about what their hopes and dreams are, but most importantly, like what the problems are. So part of this job is representing not necessarily PICI or the sites, it’s representing the network, this effect that’s happening. So if there’s a problem that we’re doing back at Central, then we need to know, I need to know, so we can go back in there and fix it and really maintain this network-based effect.

Grant Belgard: What role does bioinformatics play in your strategy?

John Connolly: Oh, it’s huge. So one of the big things with the big opportunities we have at the Parker Institute is really to become kind of a central database for all of this scientific and clinical data across the network. And it’s important that it’s written into the master collaborative agreement that knits these centers together. And so in doing so, we’ve really collected some amazing cohorts that are out there. We have the world’s largest cohort on adverse events in checkpoint blockade, just looking longitudinally at thousands of patients that are treated with checkpoint blockade. We had our RADIOHEAD cohort. And then we just published very recently with Mike Angell at Stanford, a large consortium on the BRUCE cohort, which is the largest collection of brain cancer spatial data that’s out there.

John Connolly: And all of these things are analyzed and we work at Central to make sure that that’s accessible to the network and that’s out there. And there’s many other cohorts that are like this across the network. And what that gives us is it gives the network an ability to dive in and to ask questions, whether it’s a target discovery question in the BRUCE cohort or better understanding of myeloid responsiveness in glioblastoma, for instance, or it’s something like we can predict adverse events or response to checkpoint blockade across multiple different clinical indications. All of this is available and they can dive in and really work closely with the informatics group, as well as their own informatics team. So we really are talking about the top centers in the world with some of the brightest informatics groups internally. So I think that’s one of the major areas.

John Connolly: The other is from this seat as a CSO is to push and encourage to get up there at these retreats and say what I was saying to you, which is we need to start turning the questions away from small questions to big questions and applying both informatics and large language models to the data that we’re generating. And if there’s something that’s missing, come and tell us. Like what is that data set? What is the problem that you want to solve and what’s the data set that’s needed to solve it? And so this, at Parker Institute, we can actually pull that together very quickly, put together a data strike for us to try to build that cohort. We can immediately fund large projects like that to get them off the ground. And these projects can be multi, multi, multi-year projects in order to answer really important questions.

Grant Belgard: When you’re working across multiple sites, what are the hardest data standardization problems you face?

John Connolly: Oh, so for sure. So this is my informatics team is pulling their hair out like every day, right? So always trying to stand, I don’t know, the hardest standardization. A lot of it’s probably, first of all, the simple stuff, right? If we’re doing clinical studies across real world hospitals, then it’s going to be data entry, coding, what you call like this chemo versus that chemo. Going in there, it’s actually kind of hard to, you have to go in and standardize that. The others are real batch effect things like blood processing effects. So we try to control all of that by providing templates to the hospitals prior for data entry.

John Connolly: We have our own red cap system and some of these cohorts they can enter immediately into so they’re familiar with the interface and they can put it in and it standardizes how they’re calling things like the over-the-counter meds that the patients have prior to therapy or the outcomes. The other is we use a lot of centralized core facilities when we put together these big cohorts and that’s, I think, really important. And it was a decision made really by Fred Ramsdell when he was CSO to do this. And I’m just completely benefiting from that when I’m looking at the quality of the single cell RNA sequencing on these cohorts is spectacular. It’s really, really good.

John Connolly: So being able to just take that first step, build the infrastructure to standardize that and saying, guys, for the proteomic analysis where everything is going to one site or the single cell sequencing, one group is running all of it and the samples are centrally archived and curated and we make sure that that’s taken care of. So yeah, I think that we try to tackle those problems. They’re the same problems I think that others have that are out there. But it’s important to do really.

Grant Belgard: How do you think about analysis readiness in real clinical trial data sets? What must be true before you trust downstream conclusions?

John Connolly: So I think one of the terms I just said before is real world. So one of the things, we’ve got a lot of cohorts and you and I have been involved in clinical translational studies where you’re doing deep analysis of say a phase two clinical trial, like this long data analysis where you get multimodal proteomics, flow cytometry, tumor biopsy, spatial, all of it and you get this incredible data set that’s just around each of those patients. The downside is that the patients in those academic medical centers that have those resources to do that, these are really highly selected patient populations. So when you then translate that out into the real world, those biomarkers, they don’t really translate well into real world settings. And so starting off and starting your large scale studies that you want to train these models on in the real world.

John Connolly: So go to set up in community hospitals and put the staff in there to actually pull those samples out. There you’re getting people that are checkpoint naive, they’ve got a job, they’re showing up at, you know, they’ve got real world problems. They are not the highly selected patient population that’ll make it into an academic medical centers clinical trials. So we’ve done that across 50 different hospitals for one of our cohorts now that the data is just, it translates much better into the real world. I think that this is, that’s a big one. But you do have data readiness problems when you start to do that, because these hospitals are not staffed to do any of this stuff, nor do they want to. So there’s not even an enthusiasm to do it.

John Connolly: So the other thing you won’t get is, you know, advance multiple biopsies and things that an academic medical center would do that your local hospital is just, if it didn’t have to do with care, they’re not doing it. So it’s, and they’re not funded to do it. So what we’re doing is, this is where we’re really putting in place resources at those hospitals to really acquire that data. And the other is to recognize them too. You know, when we publish these papers, we make sure that those physicians are authors on those papers, that this is moving forward. When we present the data, it’s always with those 50 hospitals in mind, the work that they did, even through COVID, to collect a lot of the data on that cohort. So, and then pulling that forward, I think, into PICI Central, where we do do a lot of QC, QA on the data sets themselves. This is, so far, this has been a, that’s a heavy process.

John Connolly: You know what I mean? This is what we hope things like AI algorithms can help with and they have to a degree, certainly for the QA they have. But, and the other, of course, is sample archiving. So, and biobanking. Always challenging when you’re doing that at scale. It’s a full-time job. We subcontract a lot of that out to CROs that I think do a pretty good job at tracking them. But, you know, there’s always the spurious sample that people are mislabeled or whatever. So, keeping on top of that, making sure you’re project managing all of it. This is, these are challenges. But I can’t emphasize this enough. The value of real world data in real hospitals is enormous to translating to something that’s actually effective in clinical setting.

Grant Belgard: What data types tend to drive the most useful decisions right now? Genomics, transcriptomics, proteomics, imaging, clinical notes, something else.

John Connolly: There’s no doubt. In the clinic right now, genomics does, of course, because we’re entering this kind of era of genomic medicine and targeted therapies. So, anything that has a matching therapy to it that’s easily available can be, that’s in the US is approved and can, you know, can be used in a doublet combination. So, this is where genomics comes in. Because everybody is getting, you know, a Caris report or a foundation report or, you know, you name it. And there’s all of this, or if you’re at some of the big academic medical centers, they have their own, you know, MSK impact or something, which does similar things. And so, with that information, you do get a bunch of really nice genomic data that can be used to guide care. And I think that that’s hugely valuable, particularly in combination with immunotherapy.

John Connolly: Right now, we’ve got, you know, tests like PD-L1 positivity in the tumor microenvironment. So, okay, well, you’ve got lung cancer, you know, it’s PD-L1 positive. We know you may, you’ve got a high likelihood of responding to checkpoint blockade. And so, that’s why, you know, it’s first line. But the questions that we want to know are not just what’s right for the median of the population, you know, but what’s right for the you personally, right? So, what is that genomic kind of workup that says, hey, maybe I should be, maybe I should start with another therapy. So, maybe I should like a MEK inhibitor or something or platinum based chemo plus, you know? I think that that’s, and then personalizing that journey is really valuable. So far, that’s really come from genomics and just sequencing for sort of driver mutations.

John Connolly: As we move forward, we’re starting to see things like, as I mentioned, what does the tumor, what does that tumor neighborhood look like, you know? So, is it PD-L1 positive? Is it rich in leukocytes? Is it, you know, what does it actually look like? Is it highly fibrotic? All of that can guide care, you know? So, and what I’m most excited about is not so much the research grade, highly multiplexed analysis of the tumor where we get huge amounts of information back. I’m really excited by the large language models that are going in and looking at just H and E stains and they’re trained on outcomes data and the genomics data that’s, or in the transcriptomics data that’s already there to actually predict, hey, this is a KRAS G12V mutant, you know, like just by looking at the H and E stain. These models are getting better and better and better at giving us more and more information.

John Connolly: And so, for me, I think I need that for immunotherapy. So, I want to, you got to move beyond just PD-L1 high, which is really just a interferon signature to not, is checkpoint blockade going to work? Yes or no? But this, your checkpoint blockade won’t work, but NK cell therapy will, or a targeted cytokine will, or an innate immune activator will. Being able to use these models to begin to predict the quality of response to immunotherapy would be absolutely, absolutely thrilling. And then from the transcriptomic side, I think there’s a lot there from the research stage, but it’s just not yet applied to medicine, you know? The biggest thing I run into is you can get a huge amount of information about these patients, but then you go and look at what therapies are available now to actually, you know, actionable therapies that are there to move on that information.

John Connolly: And there’s nothing there, or there’s maybe just one thing that’s available. So, the Parker Institute, our push, if you look at the companies we spin off, they’re a vast majority of therapeutics companies. You know, there’s like very few that are in sort of the diagnostic space. And I think that’s really because that’s where the need is. We just need more options and opportunities based on the huge amount of data that’s really coming out and a deeper understanding of that tumor.

Grant Belgard: What’s a common trap in translational interpretation where people overreach from interesting biology to clinical claims?

John Connolly: So, one is that the last thing you said, which is interesting, biology. You become just enamored with the cool idea, right? So, these ideas, I don’t wanna bash on the NK cell guys, but I’m just gonna use that as an example, that this really works incredibly well in the mouse models that we’re working on, and even in some of the tumor organoid systems or PDX models. But when we then go to apply that to someone that’s actually gone through three rounds of chemo and their immune systems beat up pretty well, and the tumor has been immunoselected to resist the immune response for the past seven years it’s been growing inside of you, this is a totally different world than a transplanted mouse or cell line that you’re trying to kill.

John Connolly: And so, this kind of dependency on clinical data, on preclinical data and this belief in it, and you become enamored with the mechanism and understanding of how it works, this leads you then to misinterpret responses on your phase one trial. So, you’re coming in, doing a safety study in phase one, you see two partial responses, you jump up and down and think this is the best thing in the world. What you missed was the eight non-responses. So, it’s one of those things that’s having, and it’s really because the mechanism works so well in preclinical data. And so, you push to the phase two, again, maybe some, and those patients are heavily selected at your academic clinical trial, so you’re just picking the patients you kind of think will work. And then, once it hits phase three, it’s 50-50 on every clinical trial, it’s a coin flip.

John Connolly: And it shouldn’t be, because you’ve been through so much to get to that point, you should have seen that this thing was gonna miss. And so, again, this is a great place for AI to come in and to remove a lot of that subjectivity from that. I think one of the most exciting things about the path I just described is, even if the drug doesn’t work, you start to, you learn a lot. There is something to be said about experimental medicine. And I think some of the best innovation in this space is really coming from clinical trials that didn’t work the way we expected them to work. It’s throughout the history of science, its serendipity has been an essential portion of it. And it continues, it’s incredibly valuable when it happens in a clinical trial, where you really say, hey, hold on a second, I’m not getting what I want, but I’m getting something here.

John Connolly: It really teaches you a lot that mice and in vitro organized just don’t.

Grant Belgard: Switching tracks now, what did you originally think your career would look like and how did reality diverge?

John Connolly: That’s a funny question. I think from a career standpoint, I’ve always wanted to be, obviously be in science. And someone had asked me a couple of years ago this, I don’t remember a time where I didn’t want to be a scientist. There’s never, never, even when I was a little kid, I had my test tubes and things like that, chemistry kits and microscope and everything in the just bedroom, even when I was six years old. So it’s always kind of gone in that direction. I think the track, as I was coming up, the traditional track would have just been an academic professor doing great things, but very siloed, where you’re doing your thing, you’re kind of in the corner of the lab, you’ve got the main project that’s driving, but you’ve got these side projects that keep you excited. I think that for me was always the track that I think people, you’d always go on.

John Connolly: I think most people that came up even now have that in mind as a potential. I think the difference really came with, as always, with the mentors that you have as you’re coming up. And I think just pointing out some great ones in Mike Fanger of Dartmouth. Mike was head of the department, but he also started a company called Medarex and a number of other companies as well. But Mike had this kind of entrepreneurial spirit, and I think that was highly inspiring. So just to watch him run the department, to do his thing, and to also run this big company at the same time, and almost seamlessly, when he’s thinking about it, he’s thinking the same way. He’s the same guy in both places. He’s just interested in solving the problem. The company, the academic position, the friends, the networks, the service that he has on profits, all of that was toward one thing.

John Connolly: And each of these different things was like Dartmouth, medical school, and the company. These were all ways to get there in service of his greater kind of vision, what he wanted to accomplish in the cancer space. And so he went on to develop CTLA-4, PD-1, all of these great breakthroughs, identified them, and then internally developed them within Medarex, which was then acquired by BMS. But I think that that, the way he worked and the way he thought was really important for me. And then moving on, I think working with Jacques Banchereau at the Baylor Institute for Immunology Research was big for me, because Jacques thought he was very much a high-energy company person. You know what I mean?

John Connolly: Getting into an environment where teamwork was essential, you had to work together with each other to get anything done, and think purely about the human disease, not so much about the mouse models, about really getting things into the clinic and testing them there. That was a change that moved me away from this idea of just the lone professor in the corner, more toward this team-based biology. And so I think those two people were highly influential. That and everybody else, you go through life. But it definitely changed the way I was thinking and the direction. You come to realize that to get anything really substantive done, you can’t do it alone. That’s all, there’s only 24 hours a day. You might be the best at everything, but there’s still only 24 hours in a day. So you really need to have a team of experts and collaborators and networks.

John Connolly: And that really led me to what I think is just a great network in the Parker Institute.

Grant Belgard: What skills turned out to be career compounding, the kind that kept paying dividends?

John Connolly: Oh, it’s mostly, it’s just the ability to work with people, you know what I mean? And enjoy it. I think some of the true, for me personally, one of the true joys in life is working with people to build things. And that for me is absolutely, it’s probably true across most careers. I said, if you like to work with really smart people, you hire people that are smarter than you and empower them and recognize them, this is really essential to building effective teams. And I don’t say that from kind of a business book. I only say that looking back on my life and what has, continues to be a really successful formula. It’s just that the ability to have joy in the successes of others and the rest of your team is really what drives this and has worked out.

Grant Belgard: What did you have to unlearn as you moved into broader leadership?

John Connolly: I think you had to unlearn the idea that you could be, I may be the best CFO, legal guy, CEO, whatever, it just doesn’t matter because again, that concept of 24 hours in a day, if I can only put four hours on that project, then it doesn’t matter if you’re the smartest guy in the room, you only put four hours on it, that’s it. And it requires 12. And so building a team of effective people that you trust, that you work well together, this is something academia does not teach you at all. I tell this story, but in my academic lab, there’s kind of no problem that’s too small for me to have an opinion on. Somebody asked me, what color paperclips? And I’d be like, it’s blue, like everything, it’s your lab, so you’re running the whole thing.

John Connolly: And that’s great because it allows you to deeply explore big ideas, but going into companies and I think probably one of the good examples when we built Tessa Therapeutics and working with a really effective team there, I knew I could walk out of a room and the right decision would be made because the people that were there, they are doing their job as well as I do my job. They are just as effective, they’re really good. And once you have that trust within the team, it just amplifies everything you can do. And again, I think that that’s something academia does not teach you. You really have to learn that in the real world or in certainly in the biotech setting.

Grant Belgard: How do you maintain scientific depth while taking on more organizational responsibility?

John Connolly: You just got to go and talk. First of all, you have to have some enthusiasm to the science, but you got to love it, right? Because you’re going to be hit with a lot of it. It’s really talking to the investigators, going out. Part of this job is to travel around to the sites to talk to them about their projects. But I don’t really want to talk to them about their projects as it relates to funding. We already gave them the funding. They have the money and they know the money’s coming. It’s about how cool the science is. And I want to kind of catch that, the excitement. And you can say it’s the young people that are going to give you the excitement. That’s not always true. These guys have been at this for 50 years that still jump out of bed and get super excited about this cool new idea. So it’s just looking at that.

John Connolly: And then from the standpoint of my own personal excitement, it’s also talking to all of those people who are ultra focused on their own projects and then knitting together those ideas. You know, like, all right, that was cool at Dana Farber, but you know, these guys at UCSF have this other thing. Maybe we should get them to talk and kind of work together. These are observations that are coming at the problem from two different angles. And so that I think is being there, talking to the people. This is what keeps you aligned.

Grant Belgard: For people who have worked across different environments, academia, industry, nonprofits, what mindset shift helps them to adapt quickly?

John Connolly: I think in academia, it’s this sort of mini monarchy that you get in your own lab. You can shift anytime you want. You know, somebody once told me, this was years ago when I was thinking about moving to a company when I was at Baylor. They said this guy named Don Capra. So shout out to Don, passed away, but he started so many great things. He’s an amazing, amazing guy. But had the good fortune to have lunch with Don. I was just asking for his advice on this company. And he said, look, academia, there’s almost no better job in the world. As long as you’re publishing papers, getting things out there, and publishing good papers every few years, getting grants in, you can study anything you want. You can wake up in the morning and suddenly decide to study turtles. And that’s what we’re gonna do. And as long as it keeps going, is there’s no job in this world that has that kind of freedom.

John Connolly: And it allows you to truly explore deep ideas. There’s huge value in that. I think from, I talked about this earlier, when it comes to the lessons from companies are very much teamwork lessons. You know what I mean? I couldn’t have done a global pivotal phase three trial in multiple countries alone. It’s ridiculous. But we could do that at Tessa with 1,500 shipping lanes and centralized cell therapy manufacturing because the expertise that we built together and the trust we had in each other. So I think that’s another one. In the nonprofit space, I think that the lessons there are that concept of mission. So as much as the lab and the company, those are very what’s going on today. You know what I mean? You’re always focused on today. Like what’s going on, we’re kind of putting fires out or keeping things moving.

John Connolly: And you’re enjoying the growth that you had when you look back on where we were and where we are now. But in PICI and I’m sure in other nonprofits, it’s really about where we wanna get to. This vision of this mission and how close are we? It’s right around the corner. And that belief is really quite essential at being effective in a nonprofit. So yeah, I think that’s kind of the take is that that’s what the nonprofit really taught me is that to have this longer term vision and to talk about mission and to constantly check back on everything on whether it’s getting us one step closer to that mission.

Grant Belgard: What does good taste look like in choosing problems?

John Connolly: Oh, wow. That’s a cool question. I think when it has to do with cancer immunotherapy, I can tell you it has to do with like, how close is this to actually curing someone? There is something else though, you know what I mean? And for me, it is that a little bit of that contrarian nature, which is, how unique is this idea? You know what I mean? Like, is it, and sometimes those ideas are, everyone’s doing this. So on that 1% chance that everyone is wrong, let’s take a deep look the other way. Like that let’s assume the sun doesn’t come up in the West. Yeah, I mean the East, it’s coming up in the West tomorrow and that’s just it. And what would that mean? How can we explore these kinds of ideas? So, and it’s absurd and the vast majority of time it doesn’t work, but it gets you to think a little bit of a different way. So for me, ideas that are audacious, I think have value.

John Connolly: And then the ones that, and when you kind of lead your way through or the ultimate kind of end point where you’re gonna get to with those ideas, if it’s leading to real effective impact in the clinic, like if this is right, then we’ve got something totally different here. Then going back and checking the data, not your data, but all the data, everything that’s out there in the field, diving deep saying, well, I can’t be right because this is going another way. So for me, it’s not following along with what everyone else is doing. Those ideas are not that exciting. Because my assumption is there are smart people that are working 24/7 on this and they’re gonna get to where that ultimate path is gonna be. But if you’re going in a different direction, you’re thinking differently, you take a big swing, that I think is an exciting idea.

Grant Belgard: How do you recommend building credibility across disciplines, especially for computational people working with wet lab and clinical teams?

John Connolly: Yeah, sure. So I mean, this team idea is essential right, but it’s also from the computational side, and this is actually even from the wet lab side, the credibility comes with the kind of end result, of course, and the track record is the end result. But to get there, and I think you know this even better than I, Grant, is you’ve gotta have an understanding. So you’ve gotta get in there and understand the biological question that you’re trying to solve, that’s all. Even if you’ve never done this before, you say, all right guys, walk me through this, let’s talk, teach me.

John Connolly: And so constantly learning, constantly engaging the wet lab scientists and learning what are the key, why is this an important question, not just what it is, to the same degree to be able to communicate, your solution to that has to be explained to the wet lab scientists as well so that they understand where this one’s coming from. It’s like, okay. And then you can really start working together, mostly because they’re gonna ask you what you think is a stupid question, but then you’re like, all right, hold on a second. There might be something there. And to the same degree, you’re gonna have this wealth of experience working with so many wet lab scientists that you’re gonna bring excellent questions to the table there too. But ultimately that credibility comes from the relationships you’re forging with these people, but also the outcome from that collaboration.

Grant Belgard: If someone wants to move towards scientific leadership, what experiences should they actively seek out?

John Connolly: So I think there’s a, I mentioned there’s a big difference in sort of this leadership role, but I think versus say your own, just a PI in the lab, there’s just so many different kinds of leadership. So one is just, is be a good leader, right? So focus on yourself. That’s the biggest thing, honestly. This idea of, I mentioned this a little bit earlier, but finding good people that are extremely talented, that are engaged, they’re energized, that share your enthusiasm for what you’re doing, that are better than you at certain areas, absolutely hire those people. And then, as I said, recognize and build. So to be a good leader, just work on yourself. That’s a huge thing. And once that happens, you’ll begin to create incredibly effective teams. And those teams will, the product of that output is really gonna kind of launch you in the right direction.

John Connolly: I think that also translates whether you’re running a small lab somewhere or you’re moving up to a dean of a medical school or you’re head of a pharmaceutical company or even just a division head within a pharmaceutical company that you’re being good leader is, people select good leaders to become good leaders. So that’s kind of how it works. And a lot of that is just focusing on your own behavior, expectations, and sort of, there’s another aspect of this, is it sounds a little Pollyanna to say that, but I think honestly, you also have to kind of filter out the noise. All the organizations I said have issues and problems and real world stresses. You just got to learn to filter that stuff out, focus on what that ultimate goal is. And this is why having a mission and about what you want to do is really important.

John Connolly: Whatever barriers come up in front of you, you’re just keeping your eye on the mission and just weaving your way around those barriers, eye on completely all the time and doing that with your team. That’s really best advice. And that’s from smallish leadership role to head of the NIH.

Grant Belgard: John, this has been fantastic. Thanks for walking us through how you think about strategy, data, and translation, and for sharing the career lessons you’ve gained along the way. For listeners who want to follow your work, where’s the best place for them to do that?

John Connolly: It’s definitely the Parker Institute website, but we also have a really active social media. So please follow everyone on Parker ICI, our Parker Institute for Cancer Immunotherapy, on Twitter, on X, on LinkedIn. There’s a good comms team. So they’re always putting out great output from the network. So, and keep an eye on the immunotherapy space too.

Grant Belgard: And is there anything you’d like to leave the audience with as a final thought?

John Connolly: There is one thing I think, I mentioned this idea of believing in the mission. You know what, I think there’s certain times in the history of science when you kind of look back on things where science doesn’t just work progressively to help society. You know what I mean? It’s not like you see like these sort of increases in lifespan over years and years and years. The way science works is inventions, people, the adoption of new procedures and technologies launch the field forward. There was a time where you were pre-antibiotics, pre-vaccines, pre-germ theory of disease, where half of kids died before the age of 11, half. And this was terrible and everybody hated it. And they felt just as bad now as then as they do now. But you couldn’t imagine a world where that didn’t happen. Well, we live in a world now where we’re living with cancer and everyone’s dying of cancer.

John Connolly: Everyone has this touch to their lives and their families and it’s a huge, huge, huge, huge impact on society. Well, we’re entering a world soon with the advances that are happening where it would be look like, it looks like childhood mortality. You’ll look back and say, how was it possible that we lived in a world where everyone died of cancer? It’s like, this just makes no sense to me. And that comes with funding this research and pushing it forward, but it is just around the corner. It truly, truly is.

Grant Belgard: John, thanks.

John Connolly: Thank you, Grant. Appreciate you.