A Hypothesis-Agnostic Approach to Accelerating Drug Discovery - Dr. Chris Gibson, Co-Founder and CEO of Recursion

Shiv Gaglani: Hi, I'm Shiv Gaglani. We've been talking a lot lately with family members of people with rare diseases about the need for more rapid development of drug therapies. Well, today we're going to take a look at a company whose mission is to create a more efficient path to drug discovery and it's using one of the most powerful supercomputers in the world to do it. I'm happy to welcome Dr. Chris Gibson, the co-founder, and CEO of Recursion, which in its nine-year history has created more than a dozen preclinical and discovery programs in diverse therapeutic areas. Chris developed the technology and approach that seeded Recursion as part of his MD Ph.D. work while at the University of Utah. After completing his Ph.D., he left medical school to launch the company, something we have in common. So, Chris, thanks for taking the time to be with us today.

Chris Gibson: Thank you so much for having me. I'm really excited to be here.

Shiv Gaglani: So, I've been a fan of Recursion ever since I heard about it from our shared investors Felicis Ventures. You obviously have a very impressive background and the more I've peeled back the layers of Recursion, the more interested I've become. So, for our audience -- many of whom want to follow in your footsteps and maybe create companies or join exciting companies -- can you tell them a bit about what got you interested in bioengineering and then drug discovery and development in the first place?

Chris Gibson: Yeah. I mean, I would have to go so far back to sixth-grade science class but I always loved science from the very beginning. In our final project at the end of that year was a thing called sludge, where we got a jar full of a bunch of different materials and our job was to use everything we learned over the year to separate those materials, figure out their density and all this sort of stuff, and then identify them. I picked the hardest sludge that one could pick. I think it had like twelve things in it, including gases that I had to capture, and it was a miserable failure. I think I got like three or four out of the twelve things, right. But ultimately, it was trying to understand something super complex like that using kind of an engineering and scientific approach that just spoke to me. My teacher actually took me to lunch, because I was pretty upset with how poorly I had done, and she just somehow turned some switch inside of me that going after these big hard things is worth doing. 

I feel like that was the moment I look back on as the start of Recursion in many ways. It started this kind of insatiable hunger for science that I've been following ever since. I tended to operate mostly at the intersection. I loved that about bioengineering. You weren't going all the way to the full depth of any engineering discipline or in biology, but trying to understand how lots of different fields interacted, and that's just always been a place that I think is a fascinating place to be.

Shiv Gaglani: That's fascinating. I love that story for many reasons, one of which is that we're an education company, at Osmosis, and many of us got into science because of those influential teachers. And you're telling me this teacher had you guys do this in sixth grade?

Chris Gibson: Sixth grade. Carol Ponganis.  I still remember her, and I think all of us have teachers like that in our lives.

Shiv Gaglani: That's awesome. Yeah, we just wrapped up this thing called the Raise the Line Faculty Awards to honor teachers like that. For me, it was Sherie Jenkins, my anatomy teacher back in ninth grade, who got me really interested in this stuff.

Chris Gibson: That's awesome.

Shiv Gaglani: Does Carol know what you're doing now? I'm curious if she's followed your career.

Chris Gibson: She does. We spoke a few years ago, and I try to make a note of following up with her every couple of years.

Shiv Gaglani: That's awesome. Well, certainly I'm gonna go more into other topics, but one last question on this sludge concept, because I find it fascinating. Did you see the recent research report by Meta where they took samples of dirt and using AI, they were able to find the structures of tons of proteins that we've never even seen before?

Chris Gibson: Yeah, I think it was like six hundred million or something like that. It's incredible. And it’s interesting to see that TikTok had a job post up recently for a digital chemist. Alphabet has Isomorphic Labs. Meta is getting into the space. There are some really interesting technical challenges that are in the field of biology that seem to be a really good fit for some of the most advanced technical computational things that humans have built over the last decade or so. So it's going to be fascinating to watch the intersection of those fields in the coming years.

Shiv Gaglani: Absolutely, yeah. All these different amazing things that are kind of the basis for a lot of the innovations that you all are leading. So, let's actually go from sixth grade to Recursion. Tell us a bit about your MD-Ph.D. and your decision about what to do the Ph.D. in, why to leave medical school to do this, and then we'll get into Recursion.

Chris Gibson: Absolutely. So, I joined the lab of a guy named Dean Li when I started my MD-Ph.D. at the University of Utah. Dean is a physician-scientist. He's actually the president of Merck research labs now, so a very translational guy who's gone on to the pharma industry. His lab brought together folks from a lot of different backgrounds to study mostly the idea of vascular stability -- how our blood vessels are leaky or not leaky -- and the role that plays in lots of different diseases. One of the diseases we were studying was called cerebral cavernous malformation, or CCM for short. It's a rare genetic disease -- actually a pretty common rare genetic disease, about six times more people than cystic fibrosis -- but one that hasn't gotten the same kind of research coverage as CF. We were trying to understand that disease as a genetic model of vascular stability and instability because people basically get...you can think of it almost like mini aneurysms in the capillaries of their brain and these aneurysms and lesions get sort of leaky. 

So, we were understanding the genetic underpinnings of that disease. This started well before I joined Dean's lab. They had built animal models, helped identify some of the genes that cause the disease and that helped dissect the molecular and cellular biology that they thought underpinned the disease, and I got to help with the end of that. We were determined to figure this thing out, and we published that it was probably driven by activation of RhoA, that was what some of the pathophysiology was driven by. So, we decided to follow that up in the mouse model and inhibit RhoA with simvastatin. Because we were an academic lab, we weren't looking to make drugs at the time through HMG-CoA reductase, and we did and we unveiled the data and we made the mice worse after five months of treatment. 

It was this humbling moment of biology for myself, and I think many people in the lab where the overwhelming complexity of this space we operate in, the space of biology, was so complex that the best kind of molecular and cellular biology tools in a lab that was regularly publishing in Nature and other high-end journals was pretty wrong. At least at the time, the data suggested we were wrong. 

That was the moment where I think maybe Recursion really started a second time, and over the coming weeks and months -- by going to various lectures, including one by Steven McKnight, from UT Southwestern -- there were a bunch of little pieces of data that got us excited in his lab about doing a phenotypic screen so that we could cast aside some of our bias about what was driving this disease and just model the genetic loss of function because we knew that incontrovertibly to be true and then ask the cells what was actually driving the disease and what could make the disease better. 

So, we built a phenotypic screen using microscopy images of human cells where we'd knocked down one of the CCM genes called CCM2 with siRNA and the cells looked radically different under a microscope. Rather than use a hypothesis-based approach, we just added thousands of known bioactive molecules and used a computer vision approach pioneered by a woman named Anne Carpenter at the Broad Institute called CellProfiler to train a very basic machine learning classifier to recognize the diseased cells and the healthy cells. 

We picked out about thirty-nine drugs, or molecules, that at the time made the computer vision algorithm think that these cells looked healthy if they had started as diseased, and we added some drugs and we went through a variety of other assays and eventually took two of those drugs into the animal model. Again, this is probably a year or two after that failure. One of them was vitamin D, and the other one was what is now called REC-994 that we have in phase 2 trials for CCM. Both of those molecules made the animals better. 

So, this traditional kind of hypothesis-based, reductionist approach led to a failure. This hypothesis-agnostic, unbiased approach led to an animal based success, and that's where we asked, could we go do this fifty times, 100 times, 1,000 times and replicate this approach to explore biology more broadly?

Shiv Gaglani: That is fascinating. Thanks for breaking that down for us. I think a lot of our learners probably have been in wet labs -- whether undergrad or high school doing research, maybe now in med school in nursing school. It's quite a big leap from their experience doing hypothesis-driven testing to the scale and the way you've used everything from in silico biology to high throughput drug screening, etc. So, can you give us a bit of a sense of Recursion as it is today? What are some of the things you are working on that get you excited and what is your scale? As I mentioned in the intro, you have one of the best supercomputers in the world, so talk to us a bit about that, too.

Chris Gibson: Happy to share. For clarity, when we were working on CCM I was using a multichannel pipette to do all that work, so it was relatively low throughput. I joke that today, we do the equivalent of my entire Ph.D.'s worth of experiments about every fifteen minutes, but that's based on this giant automated laboratory that we built. It is just a hundred feet away from me now, full of robots where we can do up to about 2.2 million experiments a week in our image-based approach. 

So, this is microscopy images. Our capacity is to generate more than ten million of those images a week across more than two point two million wells. Of course, it takes several people to operate that lab. We actually run two shifts, six days a week. But still, for a relatively small number, about a dozen people, to be running that lab and generating that scale of data is pretty incredible. 

But we've moved beyond just that image-based approach as you mentioned. We now have built our own small molecule library at Recursion, where we have nearly two million molecules physically in-house. That's kind of the same scale as many of the large pharma companies and we can screen these molecules in various disease states. We operate one of the 500 fastest supercomputers in the world, called BioHive-1. We use that to train neural nets on all of this data that we generate, and there are some specific technical reasons why we do that on prem instead of in the cloud for our training of our neural nets related to kind of the size of the images that we use. 

We also have started scaling other areas of biology. We do transcriptomics experiments at Recursion to the tune of about 13,000 of those samples a week. We have a vivarium where we do animal work, and that rodent work is done with cameras in the cages that allow us to use machine learning to visualize how the animal is interacting with its environment, interacting with its neighbors, interacting with molecules that we might use to try and ameliorate a disease model. So, really, everything at Recursion is about scale and then using technology to measure really, really broadly these high dimensional signals and to get away from the traditional western blot where you're kind of measuring one thing and it either goes up, it goes down or it stays the same. You can trick yourself a lot of ways with those low dimensional readouts when you're measuring thousands of features in biology and chemistry. I think it's a lot harder, as long as you do the statistics right, to convince yourself and trick yourself that you're accidentally moving so many different features in the right direction. 

So, we really do operate at massive scale, and today, we've generated almost twenty petabytes of proprietary biological data. To give you a sense of scale, if you took every feature length film in human history in every language in 1080p, we would have about ten times that amount of data here at Recursion. So, it's a lot of data.

Shiv Gaglani: That's incredible. I love how you communicate it because it's very analogous to what built us at Osmosis into the largest health education channel on YouTube... a lot of analogies and, you know, the fifteen Minute Ph.D. is very interesting. I'm sure some of our listeners who are doing their MD-PhDs wish that they could do their Ph.D. in fifteen minutes too. (laughs)

Chris Gibson: (laughs) I wish I could, too.

Shiv Gaglani: So, I know Illumina just announced that they have $200 whole genome sequencing available and are now trying to get to that $100 mark. The challenge becomes finding the signal from all the noise and taking these petabytes of data and finding those things that could work. Talk us through how do you go from high throughput to then actual clinical trials? And what is Recursion’s involvement once you get to, phase 1, phase 2, phase 3? Do you hand off these promising targets and promising molecules to like a larger pharma company to take through completion? How do you see it?

Chris Gibson: Yeah, great question. When we started at Recursion, we thought we could maybe just help identify these targets and then hand them off to folks. It turns out that the way the industry is set up, a couple of Ph.D. students and their professor in Salt Lake City suggesting targets does not create a massive audience of folks willing to pay the money to go advance those targets. So, we had to keep building, and we keep kind of finding that, as we build a little bit more, we start to see that there are these relationships across datasets as we go from target discovery to hit discovery to lead optimization and then preclinical testing, and now even in the clinic.  Today, Recursion has five programs that are in clinical development, four that are actively in clinical trials, and we run those ourselves. Most of those are focused in rare genetic diseases, or narrow niche areas of oncology. 

I think what we've found is that taking this tech first philosophy we've built at Recursion requires a different way of thinking in sort of every piece of the journey and because there are so few companies that are thinking the same way, it has made the most sense for us to really build our own pipeline and build the team that allows us to take things all the way through. What that allows us to do is innovate across every step, and that's both expensive and hard, but also if you can do it well, it creates a lot of opportunity. So, that's kind of where we've been building now. 

We also have partnerships. We have a partnership in neuroscience and one in oncology indication with Roche Genentech. It's one of the largest discovery collaborations ever signed in biopharma. We also have a partnership with Bayer in fibrosis. In these deals, we are actually working with our partners to uncover new targets using our technology and their teams and their technology, and then we'll co-develop those to a certain point, at which point they'll take it through the clinic. Those partnerships are around really big expensive areas of biology where a clinical trial might cost hundreds of millions of dollars. In neuroscience, think of the big ones -- Parkinson's, Alzheimer's, ALS -- these are huge clinical trials that a company like us probably couldn't do at any real scale, and you need so much expertise that we wanted a partner there. We'll focus on the rare indications that I think we can build, and then over the next decade as we continue these partnerships, we're going to learn so much from these other companies that maybe one day we'll be able to go after the bigger diseases as well.

Shiv Gaglani: That's really wonderful to hear. You mentioned rare, so let's get into that a bit. We've been doing a lot of coverage of the rare disease space for several reasons. One of the major ones is that the Orphan Drug Act is coming up on its fortieth anniversary next year, as you probably well aware. And over those forty years, there have been about 850 indications or success stories of drugs that have been repurposed or developed for various orphan indications. I had John Crowley on the podcast recently -- you may know him from Amicus Therapeutics -- and we talked a lot about this next ten years. What does it mean from year forty to year fifty with technologies that companies like Recursion have built in terms of accelerating that drug discovery? Talk to us a bit about your focus on rare diseases and why you think this is a fundamentally different era in which we can maybe see those 850 therapies go to 1,000 or 5,000.  Where do you think it'll be in ten years?

Chris Gibson: Well, I think we're gonna see exponential growth in this space over the next decade. We like focusing our work initially in rare diseases because many of these have genetically defined conditions where we know almost incontrovertibly -- thanks to that revolution in sequencing in the early 2000s -- that a mutation in gene X leads to disease Y, and that's pretty well worked out for thousands of diseases. Because of that, for us taking this target- agnostic approach, we can use now CRISPR instead of siRNA to basically rebuild those same kinds of mutations or mutation effects in human cells and then we can just watch and see what the cell does. We can explore now millions of potential molecules in that context and ask, do any of these molecules give us a rescue of this complex signature that we get when we break this gene? So, it allows us to really isolate a lot of variables and that's why we like genetic disease as a focus area for Recursion. And beyond the technical side, also because there's so much unmet need in the space as you mentioned, so that's an area of focus for us. 

Oncology also kind of flows in this same direction. There's a lot of genetic drivers of oncology that fit really well with our CRISPR-based tools at Recursion.  All of this is to say, though, that there's not just this technology revolution that Recursion is leading, but there's all kinds of new tools and technologies being built -- things like RNA medicines, for example -- I feel like all of these converging over the next several years is absolutely going to put us on a trajectory to have a really significant increase in the number of new medicines.

What's more, I think many common diseases are really going to be defined in narrower and narrower ways.  We're going to realize that one biomarker is the driver of maybe a response or non-response in a specific oncology indication that today is thought of as one big indication, but maybe it's actually several smaller ones. It's really moving us back to this idea of precision medicine, and eventually I think, really to personalized medicine which was a topic that got thrown around a lot more five or six years ago, but I think is where we're going to end up as we actually understand the complexity of biology that's around us.

Shiv Gaglani: Totally. That makes perfect sense and we've even seen that happen in real time for anybody who's been in the healthcare space for a while. The BRCA gene in breast cancer is just one of many examples. Earlier today, I had Matt Wilsey from the Grace Science Foundation on the podcast. 

Chris Gibson: Awesome.

Shiv Gaglani: You may know him. 

Chris Gibson: I do.

Shiv Gaglani: His daughter Grace, who is thirteen, has NGLY1 deficiency. I'm thinking about him. I'm thinking about people like Nick Sireau in the UK whose kids have Alkaptonuria. Actually, Alkaptonuria, you probably know, was one of the first if not the first disease that showed this one-to-one between a mutated inherited gene and then this devastating one in 500,000 condition. You know, there are hundreds, thousands, millions of these patients who have these scary rare genetic disease diagnoses. How do you decide what to work on? You only have a finite amount of time. You've obviously doing high throughput, but I'm sure patient advocacy groups are knocking on your door every day asking, "Hey, do you have anything for this gene mutation? Or that?" How do you decide?

Chris Gibson: Yeah, absolutely, and that really is how things have been for the last several years. People have been knocking on the door, and in many cases -- as was the case with NGLY1, and Matt and the Grace Science Foundation -- we did work on that disease for several years. It's a really, really tough one and unfortunately, we were not able to crack it at the time. But we've sort of moved away from focusing on one disease at a time. As we started exploring a hundred and then a few hundred different genes -- and eventually, more than a thousand genes -- we actually saw the opportunity to do what we call mapping and navigating.  

So, we've now knocked out every gene in the human genome in multiple human cell types, because we built all of the scale here to do these millions of experiments a week. And we've profiled them with imaging and increasingly with other kinds of high dimensional signatures. What that's allowing us to do is train neural networks that are trying to understand how every gene is related to every other gene in the genome, how each protein is related to every other protein, how a molecule might not interact with just one disease context at a time, but many, and that mapping and navigating has become the focus of Recursion. 

It took a long time, but in the summer of 2020, we finally got to the point where we were able to use this map to actually start to predict drugs that we thought might work in the context of a certain genetic condition without ever directly testing that drug against that gene. This is important because if you take a million molecules and you want to test them against, let's say, every gene in the genome...if you did that experiment with three replicates at three doses, even at the 2.2 million experiments a week that we're doing, it would take you hundreds of years. There's this combinatorial explosion. So, we wanted to really focus ourselves on building technical capabilities that would allow us to profile millions of molecules, thousands of genetic contexts, maybe in many different human cell types and then start predicting how all of those things interact. 

We saw that as a way, where when a patient group called us, we might be able to just look in this map and give them guidance, or look in this map and identify opportunities that we could act against at a higher scale. And that's what we've been up to for the last couple of years. I'm happy to report that our very first program that used a map to find this interaction is moving into clinical trials. So, we're starting to see the fruits of this mapping and navigating that I hope will be a huge part of what's helps us go from that 850 that you talked about before, to eventually one day having it be super normal for not just us but many companies to take “n of one” diseases and be able to generate a hypothesis and test it quickly so that we can try to affect patients that have really, really rare diseases as well as patients who have really common ones.

Shiv Gaglani: That's incredible. It's such an exciting time and totally changes what the role of every stakeholder in the healthcare system would be, from patients all the way to pharma companies and health systems.  It actually reminds me, there's a Nature paper that just came out last month -- Chris Cheadle, one of my colleagues at Elsevier was co-author on it -- entitled Old Drugs, New Tricks: Leveraging Known Compounds to Disrupt Coronavirus-induced Cytokine Storm. This was an example of the in-silico biology you were talking about where they did a screen of over 5,000 compounds and predicted that dexamethasone would be one of them that would help with reducing the cytokine storm from acute respiratory illness from COVID 19. So, we're starting to see many, many more common uses of technologies like the one you described. How do you stay on top of everything at Recursion? How do you make sure that - whether it's CRISPR, or these neural networks you've trained – they are still the cutting edge because this thing's moving so quickly. 

Chris Gibson: That's right. Well, I think we've built a learning and innovation culture. We have almost 500 people at the company. They all have different interests. They all read different articles. A group of them published in April 2020 on 1,700 molecules we tested against live SARS-CoV-2 virus in human cells, and we're seven for eight on predicting clinical trial outcomes. The only one we got wrong was dexamethasone, but we got seven of the other ones right. That driven by a set of really passionate folks who wanted to make a difference in that context and we made all that data available to the public back in in 2020. 

I think this idea, generally, of having a great team, giving them freedom to go explore, and then having a culture where people feel free to share when they have a good idea is the right approach. Good ideas rule so it doesn't matter if you're an RA or you're a scientist or you're the CEO, good ideas rule and bad ideas should be smacked down. So, we try to create that culture. It's never perfect, but we try to create that culture where our next best innovation can come from anywhere in the company and any paper anybody reads. I think we've done pretty well there. We're not always on the front of every single innovation. I don't think we could afford to be. But we ended up being near where we need to be on some of the most important revolutions in science.

Shiv Gaglani: It's amazing, really amazing. You know, before we started recording this, we were talking about skiing and snowboarding in Utah. You're in Utah. I live in Utah, too. The reason I even mention this is that I grew up in Florida and so I should be a better surfer than I am a snowboarder, but I'm way better at snowboarding and I was thinking about that. I gave a talk to Elsevier's technology team about innovation and how we do that at Osmosis, and the analogy I gave was wakeboarding versus surfing versus snowboarding where it essentially boils down to the compound interest formula: the faster you can iterate the faster you can learn.  When you’re on your snowboard in one hour of learning, you can iterate sixty, eighty, ninety times because you can get right up after you fall.  You may get a bruised butt or wrists or whatever, but you'll learn quickly. Whereas in surfing, you have to wait for the wave to come and propel you, which is every couple of minutes. 

So, what you've done at Recursion is truly like hyper snowboarding in terms of how fast you're learning and iterating as long as you have ways to pick out the signal from the noise.  By the way, since Elsevier found that dexamethasone one, maybe there's some collaboration there with what they have and what you've done to get eight out of eight as an example.

Chris Gibson: Yeah, no, exactly. And I think you're right. I love iteration and virtuous cycles of learning as a core principle for building anything that can keep up with the technology environment around us.

Shiv Gaglani: Totally. I only have two other questions. Obviously, I could make this a several-hour long podcast, but we try to keep it short for our learners who are busy between clinic and lab. I'm curious, what advice would you give to someone, say a younger Chris Gibson, who is in their MD-Ph.D. program about approaching their career in healthcare and research, moving forward?

Chris Gibson: I would say the most exciting things that have happened in my life with respect to science have always happened at the interface and it doesn't mean you don't have to go deep, but find people who have a different background than you. You know, I sat next to a geneticist on one side and a physician on the other side in Dean's lab, and getting their perspective on the work I was doing and learning about their work...it kind of created these connections that I think really were formative in what we built at Recursion and in my own career. Maybe it's my own bias as a bioengineer, but I really recommend the idea of trying to speak many languages, right? We don't do it so well in America when it comes to actual languages compared to all our friends around the world. But in terms of science, I think there are many folks who are trained really well here and around the world to operate at the interface, and that's the best place that one can be if you want to really be able to build and create.

Shiv Gaglani: I love that. Yeah, innovation is often just connecting the dots between existing ideas. So that's great. Good advice. And then the last thing is, is there anything else you want our learners to know about you, Recursion, the industry as a whole before we let you go for the day?

Chris Gibson: Well, I think one of the things that's important for folks to know is that this industry is full of incredible scientists who dedicate their lives to finding medicines. What I was fascinated to learn as I started hiring chemists and biologists who had worked for decades in large pharma companies, is the failure rate is so much higher than I ever even appreciated. About 96% of programs that start in biopharma ultimately fail before they make it to patients on the market. More if you start with really, really early discovery work. So, there are incredible, really capable scientists who will spend their whole career in this industry and end up never being part of a drug that makes it onto the market. 

I think with all the technology tools we have today, I'm excited to be a part of helping to improve that. But I also think it's important for all of us to recognize the shoulders we stand on...all the incredible publications and all the great scientists who've toiled away for the past fifty years or so to create the opportunity we have today for all of us to take advantage of this and keep driving. So, I give a lot of credit to everybody who came before us.  And I’d also just say the industry gets a bad rap, I think, because of a few bad actors and that's well deserved. But there are a lot of great folks in biopharma and for those in academics, don't totally rule it off your list. You know, go engage a little bit. Academics are great -- I almost did that as well permanently -- but there's a lot to offer here in industry and it can be very, very fulfilling.

Shiv Gaglani: Certainly. I can definitely see that. Clearly, you're still in the game, but you have some big shoulders that others will be standing on, or currently are standing on, as they continue driving this innovation. So, Chris, thanks so much for taking the time to be with us today but more importantly for the work you and Recursion are doing to raise the line and strengthen our healthcare system by making precision medicine and drug discovery and development a reality.

Chris Gibson: Huge thanks, Shiv. I appreciate it.

Shiv Gaglani: And with that, I'm Shiv Gaglani. Thank you to our audience for checking out today's show and remember to do your part to raise the line and strengthen our healthcare system. We're all in this together. Take care