BioTech Nation ... with Dr. Moira Gunn

Dr. Chris Gibson, the co-founder and CEO of Recursion Pharmaceuticals in Salt Lake City. We talk about their efforts to map ALL of human biology, and how this mapping will lead to better drugs. 

What is BioTech Nation ... with Dr. Moira Gunn?

Welcome to BIOTECH NATION !!! With understandable interviews requiring no background in science, BTN attracts a wide global audience. From everyday people looking for hope in treatments in development, to bioentrepreneurs interested in the experience of their fellow travelers, to venture capitalists looking for possibilities in cutting-edge breakthroughs, to scientists simply interested in the work of others, BioTech Nation is the voice of human endeavor, driving science to new realities for everyone. These interviews are drawn directly from the public radio program, "Tech Nation", which also can be heard in numerous global radio and podcasting venues.

Dr. Moira Gunn:

Today on Tech Nation, doctor Chris Gibson, the cofounder and CEO of Recursion Pharmaceuticals in Salt Lake City. He talks about how technology has massively accelerated the ability to perform laboratory experiments, which in turn enables recursion's quest to map all of biology. Through alliances with such pharmaceutical giants as Roche Genentech, which focuses in the neuro area, Recursion is also developing dozens of drugs in many areas. We'll talk about why their quest for insight into biology is important as well as one of their drugs, now in advanced development. The medical condition is called FAL for short.

Dr. Moira Gunn:

And those with this genetic disease develop 100, if not 1,000, of precancerous polyps in their colons. And now, Chris Gibson. Chris, welcome to the program.

Dr. Chris Gibson:

Thanks for having me.

Dr. Moira Gunn:

You said something when we were preparing for this interview that really struck me. For all that we know today and the centuries of science that have preceded us, we only know 2% of biology?

Dr. Chris Gibson:

Yeah, something like that. I don't know if it's 2% or 3%, but it feels to me like we only know a little tiny piece of what makes each of us tick inside.

Dr. Moira Gunn:

And what does that mean for drug development?

Dr. Chris Gibson:

Well, the reality is is that 90% of drugs that go into clinical trials fail. And I think that's a sign about how little we really know. Many scientists all around the world are doing their best work, and it's amazing to see some of the great medicines that we do have. But the reality is still 90% fail in clinical trials. And I think that's the testament to the complexity of biology to us that we have to unravel and decode.

Dr. Moira Gunn:

That makes sense. We only know 2 or 3% of biology, so we gotta fail a lot to get the right answer. Right. I'm with that. That makes sense.

Dr. Moira Gunn:

That makes sense. Now, of course, the technology enables what we can know. It enables the science. How has technology moved from the time, let's say, you were in graduate school until now? Describe that progress for us in terms of science.

Dr. Chris Gibson:

Well, science moves at an incredible pace. And when I was in grad school, I was doing a lot of experiments myself, so I used my own hands and the thing we call a pipette to move little drops of liquid around, back and forth, back and forth. And since then, there have been some huge developments. I think one of the most notable ones is CRISPR. So now we can use what we call molecular scissors to actually go in and cut little pieces of the DNA out of individual cells.

Dr. Chris Gibson:

And that lets us explore biology. And there's been a huge influx of new automation tools, robots that allow us to do now at recursion the equivalent of my entire PhD every 15 minutes, which is kind of humbling to think back just about 10, 12 years ago, how far things have come.

Dr. Moira Gunn:

I hesitate to say it cheapens your PhD, but, you know Yeah.

Dr. Chris Gibson:

It may.

Dr. Moira Gunn:

It feels that way.

Dr. Chris Gibson:

Feels that way. That's right.

Dr. Moira Gunn:

Now everybody knows about test tubes. You know, it's like, oh, you're pouring things into test tubes and you're looking at the test tubes and then you take Yeah. I think of it as a little

Dr. Chris Gibson:

tiny test tube, and there's a

Dr. Moira Gunn:

whole bunch of them.

Dr. Chris Gibson:

Yeah. I think of it as a little tiny test tube, and there's a whole bunch of them on something the size of a 3 by 5 index card. And the reason they're so small is that we wanna use less of the expensive liquids that let us do the science and ask the questions and use less of the cells. And so on 13 by 5 index card, just to kind of create a mental image of this, there's over 1500 of these mini test tubes on one of those plates that we call it. So each one of those is a well so that we do huge huge quantities of experimentation.

Dr. Chris Gibson:

So we do about 2,200,000 of those little wells worth of experiments every week here at Recursion.

Dr. Moira Gunn:

Well, that makes the 15 minutes look pretty busy now. So

Dr. Chris Gibson:

It it makes it feel better.

Dr. Moira Gunn:

We just upgraded you. We upgraded you.

Dr. Chris Gibson:

That's right.

Dr. Moira Gunn:

Well, now we're in Recursion's neighborhood. You've got many robots doing

Dr. Chris Gibson:

many experiments. So let

Dr. Moira Gunn:

me ask you this. What can many experiments. So let me ask you this, what can many robots doing many experiments do? Are the experiments related? Are they independent, but the data they generate brings them together?

Dr. Moira Gunn:

How do we think about this?

Dr. Chris Gibson:

Well, in the past, Moira, experiments were really done to ask very specific questions at pretty low scale. So a scientist like me back in grad school would move individual liquids around, and I might do 10, 20, 30 experiments at a time to try and answer one specific question. Today, we have robots that are doing as I shared, up to 2,200,000 experiments each week and every one of these robots is really specialized. So one of them is really good at opening the plates and closing the plates. Another robot is really good at moving the plates from one machine to another, and another robot is really good at moving liquids in and out.

Dr. Chris Gibson:

And altogether, they sort of create this symphony of activity. And the end result is this huge quantity of experiments that are sometimes related and sometimes unrelated within a given week. Sometimes we're trying to answer a specific question, but because we're so careful because these robots are so precise, all of the data we generate each week actually adds to all the data we've generated over the past 5 or 6 years at the company, and it creates one giant data set where we can actually start to see relationships between experiments we did today and an experiment we did 2 or 3 years ago. And that's really exciting because it's starting to help us unravel and decode maybe beyond this small percentage of biology that we know today.

Dr. Moira Gunn:

Now I wanna back up a bit because, I really think it's fascinating how you're able to look into these itty bitty test tubes. Oh, wells. With little things in it. Now for a single cell, which is very tiny, you put a fluorescent tag of a different color on each layer of the cell. Starting on the inside with the nucleus DNA and going out each layer to the cell membrane.

Dr. Moira Gunn:

And you shine a light on it, the computer shines a light on it, the robot does, but as you change the wavelength the different tags light up. So you can make 7 images of each cell from the inside out. Did I get that right?

Dr. Chris Gibson:

Yeah. That's right. And and you can imagine this almost like if you've ever been in a room with a dark light, and you have, like, a white shirt and all of a sudden it glows. So we're tagging different pieces of the cell with these dyes that glow when you shine different kinds of light on it. And so when we take a picture with each kind of light, we see a different piece of the cell.

Dr. Chris Gibson:

The 7 layers as you mentioned. And at the end, we've taken these 7 pictures and we see little pieces of the cell, what we call cellular organelles and that sounds a lot like organs in a human and you can think of it the same way. It's the it's the components that make the cell tick.

Dr. Moira Gunn:

Okay. So let's say that we have a cell and it's it's just by itself, we took a picture of it, we know what it looks like, and then we put our little molecule, our drug, into the cell. How might the lights change?

Dr. Chris Gibson:

They can change in thousands of different ways. Some of those dyes might, be staining something that gets bigger, maybe a piece of the cell gets bigger, or the shape changes, or it moves where it exists in the cell, maybe closer to the edge or closer to the center. And when you stay in all these different 7 organelles, and then you look at all those changes related to each other, you get an exquisite fingerprint of biological state, and it's actually not that different from what your iPhone does or your phone does when it looks at your face. It can tell all of us apart by the eyes, the nose, the ears, the mouth, the shape, the size, the distribution of those features. And we're doing the same thing with human cells, but we're doing it at extraordinary scale.

Dr. Chris Gibson:

We've taken over a 1000000000 of these images.

Dr. Moira Gunn:

Now I have to tell you that if I look back over almost a 1000 biotech interviews, lots of people have said to me, well, we put our drug in and it turned out it it would get into the cell, but it would be turned around and rejected by the cell. And then other people would say never got in. And then other people said, yeah. It got all the way in, but it just didn't do what we thought it was gonna do. Are you able to tell that with these images?

Dr. Chris Gibson:

That's right. And I think this is what's so cool about this approach is that we're not just building one experiment to ask, does the drug get in? And another experiment to ask, does it bind with the one protein we expect it to bind to? We're doing an experiment that gives us, sort of, like, a holistic representation of everything happening in the cell. It's like a map of what's happening in the cell, and so we can often tell whether a drug got in or not.

Dr. Chris Gibson:

We can often tell whether a drug is hitting different kinds of proteins. And all of that data is in each one of these images. And the more images we have, the better we can get at recognizing what certain signals mean.

Dr. Moira Gunn:

Now sometimes do you put 2 or more different kinds of cells in one of these wells for an experiment?

Dr. Chris Gibson:

Yeah. That's right. So cells interact with each other, and so you can add different kinds of cells to the well and see how they interact with each other and then see how, for example, if you add a potential drug, does it affect one cell, both cells? Does it make one cell affect the other cell? And this starts to help us further unwind that complexity of biology that we've been talking about.

Dr. Moira Gunn:

So let's put it together. You have 7 images of the 7 layers of a cell all by itself, and then you have 7 images of the 7 layers of a cell when it has a drug and or 7 images of 7 layers of a cell with different kinds of cells. This means a lot of images and a lot of data. No human could possibly look at all these images and data and understand it.

Dr. Chris Gibson:

That's absolutely right. I mean, we're talking about more data than every film in human history in every language in full 10 80 high definition. So, like, I don't know anyone that's watched every movie ever created, but that's the scale of the dataset that we're operating with. And to your point, it's way too much for anyone. I mean, when during my PhD, I felt overwhelmed by the amount of data I was generating.

Dr. Chris Gibson:

And as I said before, we generate that every 15 minutes here. And so we have to use computational tools to help us not only sort through that data, but actually what's amazing, Moira, is the computer algorithms that we've built can see things that actually no human can see even if we had the time to go through all the images.

Dr. Moira Gunn:

Now let's talk about what recursion does. From my perspective here, it's 2 things. You use this capability to develop drugs, that you do that yourself. Recursion does that. But you also are in partnership with other very well known drug developers.

Dr. Moira Gunn:

And let's start there. As I understand it, you have a a 10 year partnership going forward with Roche Genentech. I guess there's also Bayer in there. What are you doing? How are you working with them to the extent that you can tell?

Dr. Chris Gibson:

Absolutely. So with Roche Genentech, we're exploring the whole of neuroscience. So all of us are familiar with terrible diseases like Alzheimer's disease, Parkinson's, ALS. And these are actually areas where there have been a few good drugs, but actually way, way fewer than I think all of us wish there'd been. And so we're going on an exciting voyage with our colleagues at Roche and Genentech to map neuroscience.

Dr. Chris Gibson:

To try and understand how neural cells from people respond not only to all these different potential drugs that we have at recursion and and at our colleagues, locations, but also to understand how every gene in the human genome actually plays a role in these different cells. And the hope, the goal, is for us to actually develop up to 40 new medicines. 40 new medicines in neuroscience and one oncology indication with our colleagues over the next decade. And that may not sound like a lot to everybody here, but there's only about 40 new drugs made every year in the whole world. So this would be like for just us and them over 10 years to do what the entire industry is doing, in a single year.

Dr. Chris Gibson:

And and not only that, doing it in an area of biology that's so so challenging.

Dr. Moira Gunn:

And recursion also has a dozen drugs in their pipeline, and I'd like to talk about the most advanced. You have 3 drugs in phase 2, which is just, you know, prior to phase 3. Let's talk about the one that relates to colon cancer. What is it? What's it trying to do?

Dr. Moira Gunn:

And if we were a subject in the trial, who would qualify? What would happen to us? Take us from day 1.

Dr. Chris Gibson:

Absolutely. So there's a type of colon cancer called familial adenovitis polyposis, and we'll just say f a p for short because that's quite a mouthful. And in this particular type of colon cancer, patients actually are born with a genetic mutation in a gene called APC. And these patients will all get colon cancer. It's a it's a really, really terrible disease.

Dr. Chris Gibson:

And so today, the gold standard treatment for a patient with this disease in the developed world is they actually have their colon removed typically in their late teens or early twenties, and they live the rest of their life without a colon, which has all kinds of effects on on somebody's, you know, ability to live their life, the quality of life. And what's more, they still have a really high chance of getting cancer in the part of the g I system that remains. And so our drug was discovered using this mapping and navigating approach that that we've talked about using these images and other tools that we're building here. And this drug, we hope, is gonna have an effect in these patients. And so the trial that we've just started is taking patients who've had their colon removed to have this disease, and they're randomized and neither they nor the physicians who are administering the drugs know what they're getting.

Dr. Chris Gibson:

So it's called a double blind placebo controlled trial, which is kind of the gold standard, and they'll be given our drug or a placebo for a year. And during that time, we will be measuring the number, size, and location of polyps inside their GI tract. And polyps, for anyone who's had a colonoscopy, polyps are these, you know, little lesions that are in your colon or in your your gastrointestinal tract that are sort of like precancers. And often in a colonoscopy, they'll actually snip these out if they look worrisome. But these patients can have 100 of these polyps, and we'll be trying to ask the question, can we stop the growth of new polyps?

Dr. Chris Gibson:

Can we actually reduce the number of polyps? Or could we actually even eliminate the polyps and especially the ones that look like they're precancerous? And if we could do that at the right kind of levels, then this would be a successful study and would be able to work with our partners at the FDA and the EMA, which is the European version, to take this drug forward to try and get it to these patients at a broader scale.

Dr. Moira Gunn:

And it would be great if you could do it before they had their colons removed.

Dr. Chris Gibson:

Yeah. That's exactly right, Laura. And I think that's where we'd wanna go next. But that's such an effective treatment today that it's actually really not something that we can do in our first trial. First, we have to show that this drug is helpful in patients after their colon's removed.

Dr. Chris Gibson:

And if we're able to demonstrate that, then we could have the discussion with patient groups and with the FDA about a more aggressive trial, which would be to see if we could delay or even eliminate one day the removal of of their colon. Because as we shared before, that's, you know, really a really big event in somebody's life.

Dr. Moira Gunn:

How many subjects are in this trial?

Dr. Chris Gibson:

So most of our trials are focused on relatively rare genetic diseases, and we'll have somewhere on the order of 60 to a 120 patients in these trials.

Dr. Moira Gunn:

So we have all this work that your robots have done, and created all of this data. And not only that, you're now going into these, you know, the full FDA trials and collecting that data as well. Can all of this be tied together with others, with other data that is being created or has been created?

Dr. Chris Gibson:

Absolutely. And we we use data that others have created. One of the challenges though is I talked about the kind of, noise of biology. And the reason we use all the different robots is because if you do, an experiment at recursion and then you do an experiment at a different lab, in a different lab, in a different city, often there are little variations there that can be really hard to understand. And so one of the real differentiators for us is that we've focused on building most of our own data sets that we use, and we build them in house.

Dr. Chris Gibson:

And that's really, hard and expensive to do, but we think it's a really important piece of ultimately building a dataset that allows us to see much more subtle connections across biology and chemistry.

Dr. Moira Gunn:

I have to give you an award here, Chris. And I say that because you said all of this. You told us all of this, and you never used the term artificial intelligence. But this really is artificial intelligence. Right?

Dr. Chris Gibson:

Well, that's exactly right. And, obviously, that's a term that's become overused, in many industries. But for us, we really have trained something called convolutional neural networks. So these are essentially programs or algorithms that are trained just like we train our brain by showing them lots of of different images and lots of different results and training them to understand and recognize differences. And as I shared earlier, they've gotten so good that they can see changes in these images of cells that we take that no human can see.

Dr. Chris Gibson:

So they're sort of better than we are. They've evolved past us at this one specific task. They don't do a lot of general stuff. They're really trained for this very, very specific task, but they far exceed our own ability at that specific task.

Dr. Moira Gunn:

Okay. Here's my last set of questions. So it's just one cluster altogether. How big are these various robots that, you know, some analyze, some load the material in, some move things around. And and how many do you have?

Dr. Chris Gibson:

Well, I have a confession to make. These robots are a little disappointing if you're a a a 6 year old or a 9 year old. Everybody expects when we say robots that they've got faces and arms and wheels and they're they're jetting around. The reality is they are large beige boxes that sit in our lab and they do cool science. But if you come look at our lab, you don't see quite as much movement as you might expect from, like, a sci fi movie.

Dr. Chris Gibson:

But we have dozens and dozens of these. So just the the robots that take the pictures of the cells. We have over 16 of these these robots and they're they're nearly half a $1,000,000 each. And each one of these robots is generating some of the images that we ultimately use. And this is the biggest installation of those kinds of microscope robots anywhere in the world.

Dr. Chris Gibson:

So it really is a substantial operation, but my 6 year old and my 9 year old are always a little disappointed because our robots aren't as cool as the one in the cartoons.

Dr. Moira Gunn:

Yeah. At dad's data, it took 15 minutes. He's very unimpressive. I don't know what to say. How many employees do you have doing all this work?

Dr. Chris Gibson:

So we're just under 500 employees now.

Dr. Moira Gunn:

Interesting. The robots are doing a lot of work, but they we still need the humans.

Dr. Chris Gibson:

Well, we absolutely do. And I think this is a really important point of the advance of of technology is that, you know, we're growing still as a company and and nearly 500 people. And about 40% of those people are biologists and chemists and about 35% are software engineers and data scientists who are helping the robots and the algorithms be directed in the right ways. What the robots and the algorithms help us do is they help that 500 people group just do more work and be more efficient and hopefully find better drugs more quickly. But there's, I think, always gonna be a huge need for us for great scientists and and great employees at at Recursion and this is true, I think, across every company in the industry.

Dr. Moira Gunn:

Well, Chris, thank you so much. You're always welcome on TechNation. You come on back anytime you'd like.

Dr. Chris Gibson:

Amazing. Thank you. This was so much fun. Appreciate it, Moira.

Dr. Moira Gunn:

Doctor Chris Gibson is the cofounder and CEO of Recursion Pharmaceuticals based in Salt Lake City. More information is available at recursion.com. That's recursion, recursion, recursion.com. I'm Maura Gunn. You're listening to Tech Nation.