Next-Generation Biotech Platforms: Engineering Biology

Welcome to the Axial podcast. Axial is an early-stage investment firm based in San Francisco. We partner with great founders and inventors investing in early-stage life science companies, often when they are no more than an idea. Axial is fanatical about helping the rare inventor who is compelled to build their own enduring business.

Hello? Hello? Hello? So weird.

Hey Brian, how are you?

Hey Josh, how's it going?

Doing pretty well, kind of did some phone calls and stuff. But first things first— oh, Nikesh is here. How you doing, Nikesh? Hey, Nikesh. One thing, first thing, welcome to Clubhouse. But if you click on my profile, you can make me a moderator, just so I can filter the crazy people.

Done.

Cool. Cool. This is like— well, awesome. How was your Thanksgiving?

Pretty good, thanks. Yeah, just relaxing. We're going to be traveling for Christmas, so we're staying home for Thanksgiving.

I don't think, I don't think, I don't think that's— I think this COVID stuff, I think they're gonna— we'll see how bad this variant is, but, uh, depending on, uh, well, I think I lost like 2% of my money today, so, uh, I'm gonna go take a little walk and deal with it. But, uh, I think we all lost 2% of our money today, uh, uh, but, um, yeah, great to have you here. It's not bad, it's not that bad actually. I have hopefully these startups go public. Hopefully Hexagon goes public one of these days. That'd be cool. That would be cool, yeah. But congrats on 23andMe. I mean, you already— that already went public. But yeah, great to have you here. I think really excited to have Brian here to just have a conversation about his career, his work, and just about biology in general. I met Brian Norton through a friend of mine, the Farlow brothers, at a an investment dinner. And so me and Brian have gotten to know each other over the years and I love hyping people up. So let me, I'm trying to hype you up as much as possible. And Brian's an incredible biologist and engineer and he's a superstar. I think more and more people should get to know his work. And I'd really recommend you read his website, Boolean Biotech. It's really, I'm a huge fan of that CGM thing you did. YouTube. That was really cool. I've never seen anybody do that so comprehensively.

Yeah, I went overboard. You did.

You went a little overboard there. So I recommend you read that great blog post. He has a lot of other great blog posts. But yeah, I think Brian, maybe you want to do a quick introduction on yourself and then we can go from there.

Sure. Yeah. So I'm Brian Nocton. I'm mainly a genomics guy. I did a PhD in biomedical informatics and then was founding scientist at 23andMe. That was, um, 2006 to 2014, something like that. Um, then I left 23andMe, spent a couple years doing sort of other things. I think that's when I met you, Josh. Um, and then co-founded Hexagon around 4 years ago with Maureen Hillenmeyer and Colin Harvey, and we've been working on using fungal genomics to discover new natural products ever since.

Cool. I think for me personally, I never asked you this question, but it'd be great to go to Trinity. So you went to Trinity College in Ireland?

That's right. Trinity College Dublin.

Dublin. Were you born in Ireland or did you go there? Okay, cool.

You were born.

How do you not have an Irish accent? Did you go to school for that?

I've lost my accent somewhat over the years, over the 20 years. But there are a lot of accents in Ireland and they don't all sound like the stereotypical Irish accent. If you're from the part of South Dublin I'm from, the accent is actually not very strong.

You have a slight, it's slight, but it's Yeah. Okay, it's very—

there's something I'm a little curious about.

I think, because I'm half British, so I had a British accent as a kid. So my, I was like in third grade and my teachers, they sent me to a class to like not have British accent. And it's like, it's horrible, right? Imagine me with a British accent, how cool I would be? Now I talk like this, so I can tell someone's curious. But seriously, that was a serious thing. I had to go to a language class for a year. And I had to learn how to say R, R. I still haven't mastered that one. Saying R is so hard, actually. It took me a year to figure it out.

Yeah, I still say or, like PCR. I'll never change that one.

But how was Trinity? Trinity is a really iconic university. I mean, Schrödinger gave his What Is Life speech there. That's right, yeah. Maybe you can discuss how you got into biology, any formative experience at Trinity and stuff like that.

Yeah, I'm— yeah, I'm happy to talk about that actually, because I feel like it was pretty important to me and they did a really good job of sort of being ahead of the curve. And, you know, it's a good university. I mean, it's a very good university, but they're not always like ahead of the curve, right? Because it's 500 years old or whatever. But this, this professor there, David McConnell, had sort of been to the US. He'd gone to Caltech and, you know, worked under— I don't know, I think he maybe worked with Kornberg at some point. So he'd sort of seen the, you know, the foundations of, of early, early biotech. And when he came back to Ireland, he set up a genetics department in Trinity. And then as part of that department, he set up a very small program called Human Genetics. Because he just wanted people to learn human genetics. So in my class, my human genetics class, there were only 8 people. And this was 1997 we started. And I mean, that was pretty early, right, for to have a specific genetics class. But he was sort of convinced that genetics was, you know, on the rise. They had sort of strong geneticists in that, in that department, and a few sort of random, pretty random people from around Ireland selected that course and got in because in Ireland you do your what's called Leaving Cert. At the end of high school, you take this big exam, you get a certain number of points, and then you rank order your preferences in, in university, and then you go into whatever matches, right? Your sort of whatever your points allow you to do. So medicine traditionally would be very high points, and then human genetics was sort of a weirdo outlier. And just a few of us decided to completely jump into the dark on that one. So actually, of the 8 of us, 4 are now in the Bay Area. Actually, that's not true. My friend was in the Bay Area. Now he's in New York working at Oxford Nanopore. Then another, uh, one of my classmates is at Zymergen, and another one of my classmates was at 10X, and now she's at some other company that I don't remember. But half the class actually ended up in the Bay Area, which I thought was kind of crazy. Um, so yeah, anyway, Trinity was, was great in that way.

That's awesome. Well, I think there's always these kind of hotbeds of talent, and they just— I'm always amazed how people self-assort. Yeah, and it's always, always in, in throughout history just these talent hubs and they all kind of, you know, so let's talk about another talent hub you're part of, maybe even bigger at Stanford. So how did you go from Ireland to Stanford? Was it kind of, was there a mentor of yours that motivated you? Was it a particular researcher you wanted to, or did you just want to go to Silicon Valley and just make your stake, your claim?

Um, sort of like a lot of other people, I wanted to do, you know, a PhD. And I was actually, again, lucky that one of the professors at Trinity in that genetics department was a guy, Ken Woolf, who was an early bioinformaticist working on yeast genomics. So there was a lot of scope to get exposure to genomics and computational biology there, which again, sort of fortunate. So I knew I wanted to do a PhD in something computational. And then I'm not sure exactly why I I thought I would apply to U.S. universities, but it seemed like the thing to do. So I applied to a few of them. UCSF and Stanford especially, I think, were at the top of the list. I think I just didn't want to go somewhere cold. And then I was very lucky to get into Stanford and, you know, said yes, why not? Right? It's like something to do. I was pretty young, so off I went.

So you're in the, if I remember correctly, the biomedical informatics program at Stanford.

Yeah.

And then, you know, to be honest, you're that crew. You have Nikesh as part of that crew too.

That's right. Yeah.

And Nikesh is obviously, I tell Nikesh all the time, he's next up. Nikesh, he's next up to build a great company. So he needs to get on that actually. And I'll write him, we'll write him the first check, right? Nikesh, feel free to raise your hand if you want. But how was grad school? You had a lot of great peers in your class. You know, any, any like kind of formative stories, experiences while at Stanford?

Yeah, I mean, just looking back at it, it was kind of crazy, right? Because, you know, there were like the BMI program, biomedical informatics program, it was pretty small, but I felt like a lot of us went on to be involved in interesting things just because At that time, if you wanted to be involved in genomics especially, there wasn't really a genomics program. There were definitely some people in the genetics department at Stanford, but a lot of the computational biology converged on this BMI program. And because it was early, it was just a few people. So then, um, yeah, there would— it just turned out that that led to a lot of, uh, companies getting started, honestly, was the main thing. So looking back at it, right, I was in Seraphim Batseglu's lab for, for a time. And out of that lab came DNA Nexus and, and Council Biology. Srinivasan was there at the time. And then Atul Bhushe and, and Ross Altman were part of the BMI program at various times, and that they started, or with a few other people, this Personalis. Steve Quake was around. I was in sort of kind of drifted into Pat Brown's lab for a very short time. He obviously left Stanford, started Impossible Foods. Um, Gardant— uh, the co-founders of Gardant were in the Genome Center at that time, which is where Maureen was. Uh, Daphne Culler obviously was there, um, teaching genomics, and Vijay Pandey was there teaching structural biology and stuff like that. Um, yeah, even Mustafa Runaghi, the former CTO of Illumina, was at the Genome Center at that same time as well. So there really were like, in retrospect, a lot of people around who would go on to sort of start companies. Serge Saxonov as well, of course, he was my, he was my labmate, and he co-founded 10x Genomics, hugely successful genomics company. So yeah, I think pretty lucky to get sort of into that crew, you know, just know those people. It was very lucky.

You also, Maureen's part of that crew too. So your co-founder for Hexagon. I remember seeing this picture, you know, Serge was having a baby shower. I can send you the picture. I don't want to dox Serge, but he's at a baby shower and there's a photo of Nikesh and Maureen. I think you're in that photo too. Maybe not actually. And so it's just something I find fascinating, all these friends early on in their scientific career. And Serge is like, he built like, how big is Tenet? Like $20 billion or something? Some huge company.

Yeah, something amazing.

It's going to be worth $100 billion pretty soon, one of these years. But it's just amazing to have all these talented folks together. And so while in grad school, what was the kind of the experience in terms of like entrepreneur— how did you learn about building companies? Was it just from professors, or, you know, was it kind of always— was it like a blood pact, hey, we're, we're going to all start companies, or something, something like that?

Well, yeah, that's, that's a good question. I think so. For Serge and I, we were both, I think, very interested in, uh, industry more than academia, and we were both in Doug Bootlag's lab, and Doug had actually been involved in a lot of startups sort of previous cycle of biotech startups. Um, I just been sort of like, yeah, part of that scene. He was very, uh, pro-industry, and, um, that just seemed to be the way that his lab was trending. People out of his lab often went to— well, one fellow, one guy went to Google, and someone else started some kind of chip company that I don't remember. So it was quite technical as a, um, as a lab, but not very, not particularly like, you know, paper-focused, like the lab wasn't publishing a huge number of high-impact papers, we were sort of like developing tools, which was very useful, I guess, for, for industry, or training for industry. And then the BMI program also, interestingly, was, was pretty industry-friendly. There had been a bunch of companies that came out of there over the previous sort of 10, 20 years. And again, it was a, it was a program that was very focused on making tools, like sort of the, one of the very first projects out of the BMI program was like diagnosing diseases using computers, right? And you can imagine over the, over the years, a bunch of companies tried to sort of solve problems in that area, like can you take electronic medical records and turn them into something useful? This was the idea. So I don't want to say it wasn't like an academic program, but it was very tool-focused and very sort of practical. And I think that's partially why so many of the people in there went on to, to start companies. I guess the other, the other factor was that, you know, the, the computer science department had seen a glut of companies come out and then we were sort of like vaguely connected to computer science, right? Like a bunch of BMI people would be in computer science labs.. And you sort of see Google started, you know, 5 years before, or Yahoo, or all these like CS companies. And that sort of starts to make, you know, the gears turn, right? It's like, well, if that happened for, you know, the internet, maybe it should happen for genomics. And what would that look like? And, you know, and that was part of it as well.

Yeah, I'm more convinced it's the latter. Just if you see somebody run up— if you see someone run a 4-minute mile, you're like, well, maybe I could do it. And so I think, yeah, in startups, you just see, oh, somebody who kind of looks like me, talks like me, did really well. Maybe I can do it. And then you should go from there. Maybe shift gears, that you should go from Stanford, and then you go to a third talent hub, and you start— you're the one who builds it though. So you go to 23andMe as the founding scientist. I don't know if— I know Serge— was Serge the first employee, or were you the first employee?

I'm always—

Yeah, Serge has the mantle, right? He's the first employee and you're the founding scientist.

Yeah, we were both founding scientists. I think that was the same title. Serge finished before I did, finished his PhD. So then there was kind of like, I don't know, a couple of months of delay between us. But yeah, definitely Serge.

Okay, so you joined 23andMe. I remember you built the fab team, you know, Farr, Aaron, and Babak. That's a very legendary talent hub you built. But, uh, uh, uh, you know, all three of those people are superstars. I think Farr's— they're all great companies now. But, uh, but what made you want to join 23andMe?

Yes, Serge and I had been discussing sort of what, what would be a fun thing to do. And like you said, other, um people coming out of Stanford were starting these companies and having huge success, and we, we wanted to, uh, to do something like that. I think it was always going to be something internet-related at that time. That was kind of like the, the thing, um, and also genomics-related. So we had been sort of trying to figure out what we could do in that, um, in that area, and these SNP chips, genotyping chips, were kind of coming out And there were papers coming out saying, you know, if you have this SNP, then you have this condition or you have this predisposition to this condition. So that was kind of like what we had settled on as something to do. And we had sort of like started assembling, you know, pitch deck, just trying to see like how it would work. And then we met Anne Wojcicki and Linda, the founders. And basically decided, you know, they had already funded it, um, had a great name, um, and it seemed like they had, you know, a good vision for what they wanted to do. Um, like specifically, Anne really wanted to do something big in healthcare and not just sort of, you know, sell things online. Um, so Serge and I joined basically with the idea that we would build the science and the technology behind it. And that's what happened. That's how it worked.

I didn't know that. Okay, that's cool. Okay, you and Serge were kind of doing your own thing and then happened since you met Anne, who is just, you know, had Google money. So, okay, might as well just join her versus compete. And then, and so what was it kind of that first year like? Because you were kind of 23andMe is definitely a category-defining company. Right before 23andMe, there's no such thing. It was like, I think Bill Burr has a joke. It's like sending your DNA through the internet. Yeah. What were the opportunities? What were the challenges? Because it definitely wasn't just a drug company. It was amorphous. And how did everyone on the founding team just wrestle with these new ideas and you know, you're definitely a regulatory— you had regulatory issues, you had, you know, consumer adoption stuff. But that first year, what was it like?

Yeah, I mean, the first year I think was not so crazy because we were just figuring out technology, right? It's like, well, which chip should we use? And there was an upstart company called Illumina at the time that was, uh, you know, 1/10 the size of the, the gorilla Affymetrix. Like, it was a small, small company. Illumina, but they had this really good chip technology and we had sort of like decided this was the thing we would use. Um, so we started working with Illumina really, really early actually, before their sequencer took off. And, um, that was really the first year. I think things got much crazier after we started to sell it, and that was— I don't know how long that was in, but the thing that I always find kind of strange or hard to, hard to grok now is how controversial it was that you would be able to get any health information online. So WebMD, I think, was somehow grandfathered in like there. You can sort of Google your symptoms and get something back. But anything that involved a test was extremely controversial. And again, it's hard to like— it's hard to believe at this stage, but And or sometimes I would like go on these radio shows and just defend the idea of being able to get your data back, defend the idea of doing genetic testing outside of a doctor's office, defend the idea that like you wouldn't— it wouldn't lead to something insane happening. Like, you know, X percent of people, you know, have the incorrect paternity, therefore it's going to cause, you know, social unrest because 5% of people or whatever the number you get is they're going to find out that their father, who they thought their father is, not their biological father, or they're going to, you know, overdose on something or whatever. And these were like not unusual things to hear for us. These were like the normal sort of worries. So again, like now it seems like the most normal thing in the world, telehealth and getting information online. But at the time it was really surprisingly controversial. And that, and we got shut down a couple of times along the way, right? We got shut down because we were doing testing like in a, in a facility that was like technically research use only. We had sort of, you know, figured that this was okay by talking to people. But, you know, regular regulations are pretty fluid, right? Even though they're written down. It's sort of often at the discretion of whoever is running that agency. So we sort of found that we had to be very flexible. Um, when something happened, we just had to figure out some other way to, you know, run the tests or, or whatever. Um, so yeah, it was, it was a pretty, pretty crazy ride. I also remember one more thing, I guess, on that topic. I also remember sort of presenting at the American Society of Human Genetics. So this was, still is kind of the big human genetics meeting for scientists and almost like getting booed out of there. I wasn't the one presenting, but they were really, really angry at the idea. And I think partially it was because these genome-wide association studies were getting published, but then nobody was really expecting that anyone would use the results. So it's kind of like you can publish something and then assume like, okay, whatever, it's just data, right? It's just out there. But then we were using the data, right? We were using those, that information to give people knowledge about their health. And that made the academic geneticists extremely angry. And it took a few years, took maybe 10 years actually before it became normalized. And now A lot of those professors started companies that do the same thing, right? Or do it in a slightly different way. But yeah, it was controversial everywhere. Interesting.

Well, that's really exciting. I mean, one thing about 23andMe is that I think early on, the adoption is driven by celebrities. Yep. Did you meet any cool celebrities early on? Did you meet like Brad Pitt or something?

I don't think he was around. There were random people coming in and out of the office. The coolest one's probably James Wilson. I met James Wilson.

He came back.

That's awesome. Spent some time with us.

The gene therapy guy, right?

Just— I'm sorry, did I say James Wilson? James Watson.

James Watson. Oh, James Watson. That's even back when you were a son. That's pretty polarizing.

Yeah, yeah. No, very controversial. Yeah, very unusual person, but, you know, still came by the office, kind of a famous character, and that was pretty cool. Eric Lander came by a bunch of times. I know he's not like exactly a celebrity, but he was sort of like, I don't know, he and Anne knew each other, and they were— so he would come by and sort of give advice sometimes. I do remember Goldie Hawn and whoever Goldie Hawn's husband were in the opposite side at some point, but I don't remember why or if I met them.

That's hilarious.

Anyway, yeah, occasionally there were, but most of the time it was just, you know, some scientists and some computers.

So you met Goldie Hawn and Kurt Russell. That's pretty— that's— those are pretty good celebrities, actually.

I don't think I met them. I think they were in the office, though.

They were like, okay, you're in the vicinity of celebrity, at least.

Are you—

as a celebrity now, anyways. Right. So for 23andMe, you're kind of doing incredible science, but you also create a whole new category, and that's kind of very rare in biotech to kind of do business model innovation and like try to do a scientific, just make technical progress. So reflecting back on that experience, what kind of key lessons have you learned? Do you think that more 23andMes can exist in the world, or was this kind of a one-off thing that happened, just the right place, right time? Anne too, right? I would say Anne played a huge role in the success of the company.

Absolutely.

Given, you know, she was able to hunker down. She had the money to hunker down, whereas maybe other founders wouldn't be able to do that. So any kind of lessons, maybe even lessons that can't be replicated, get?

Yeah, I, I think there's definitely room for more, you know, DTC health, um, companies. I think that's just getting started. I think things like Butterfly are a good, good example of that, right? It's sort of like, um, this is an ultrasound device, it's portable, it's not really DTC yet, but I think that's the kind of thing where I would love to see a lot more of that and it just get normalized. I think 23andMe did have to to honestly break down those— a lot of those barriers. And I think it was an expensive thing to do. So like you said, and was extremely important for just having that persistence and keeping at it and not sort of like pivoting to, you know, providing through doctors, right? Because a lot of the other companies that came up at that time were sort of direct-to-consumer, but not really direct-to-consumer. They would sort of act like direct-to-consumer companies, but you'd still have to get your doctor involved. And we were really sort of adamant, or really Anne was adamant, that we would be truly direct-to-consumer. And I think that, that really has made a big difference. And I think hopefully other companies won't have to go through that sort of expense. It'll be easier to go sort of straight to selling to people that need whatever it is you're selling.

Absolutely. I mean, they have a bunch of copycats and so 23andMe definitely just did all the hard work to show like, hey, this is really possible. And then 23andMe kind of added on drug development capabilities and then congrats on going public. That must have been a really exciting day. That's cool. Yeah. But maybe we could shift gears towards kind of the meat of the conversation. You know, what you're building at Hexagon and how you made that transition. You're at 23andMe doing all this work. How did the, how did, you know, even Hexagon get started? Was it just running into Maureen? Was it, yeah, how did that go? Because it seemed like it was a kind of a, maybe a year or two gap in between. Yeah. 23andMe and Hexagon. So how did, how did that process go?

Yeah, so I left 23andMe and then I, I sort of knew I wanted to do something a little bit different to diagnostics. So at the time, diagnostics was really not a great field, even though it wasn't that long ago. Um, it kind of seemed like you had to do either non-invasive prenatal testing or liquid biopsy. Or something that's a lot like 23andMe. Those were kind of like the options for, um, you know, and I think they still probably are the options for a DNA-ish diagnostics company. So I really wanted to do something in therapeutics, um, but somehow using genomics because that's, you know, what I know. Um, so I left— when I left 23andMe, I didn't have another job, but I did sort of want to do do some experiments and try and just see what's going on, like put my head up above the, um, above the water and have a look around, because things had changed pretty fast, right? When we started 23andMe, um, everything was chips, so it really had been a long time. And obviously things had switched pretty fast to sequencing in those years. So, um, I started writing that, the blog that you mentioned. That was kind of part of it, just sort of in order to force myself to, you know, write things down and say what I was doing. I felt like that was a useful exercise. And through that, actually, I met a lot of interesting people. Like, I met Justin Farlow from, from Serenity, and then you. Um, through that process, I met, um, you know, Rich Stoner at Synthego through that blog, and a bunch of kind of people that were thinking along the same lines— how do we automate biology How do we use genomics for therapeutics? How do we scale some of these problems, um, to do more things, explore more proteins, do more sequencing, whatever it was? Um, so anyway, that was a very useful sort of year or two. Um, I also did a project with Jake Glanville. So, um, got to know Jake through Nikesh, actually. Nikesh introduced us. So Jake Glanville, for those who don't know, is, uh, CEO of Centivax now, but he used to be CEO of Distributed Bio, antibody company. So I'd been working with Jake for a little bit during that time to try and develop an antibody library, and that actually worked out pretty well. That was a fairly successful endeavor that we, that we started. But it was hard, honestly, to do all that stuff by myself. It was just like, uh, it was sort of difficult to, to be like a one-man show a little bit. So then I started working with Maureen and Colin Harvey at Stanford as sort of computational biology consulting, sort of to figure out what they were doing, see if there was anywhere for me to fit in. And around that time, they had just gotten this heterologous expression platform working. So for those who don't know, heterologous expression means You take the genes from one organism and you put them in yeast, and then you get the yeast to express those proteins. So you get it expressed in a different host. And they had had a hugely successful project there to get that working for fungal genomes in yeast. And then it became pretty clear sort of through the process of working with them that there was a prospect for a company here. And basically, Within a few months, we had decided this was the thing to do and started pitching.

Yeah, really cool. I remember reading the paper, like the HEX paper, like H-E-X, I think Colin, Maureen, Pablo, actually Pablo was on the paper too, if I remember correctly. And so they had this work at Stanford to express biosynthetic clusters in yeast and make all these natural products, and then you were helping them kind of build out the data capabilities. Um, what was the kind of the moment that like, hey, uh, um, this is a company, we're gonna start this? Was it— was already a company when you joined, or was it kind of like, uh, or was this a project?

It was really a project, I think, when I joined. I'm certainly, um, Maureen had been interested in startups around that time. So I think she was open to the idea. It's not like they, they had no designs on doing a startup and then I said, hey, you should do a startup. I think it was kind of natural. And then I think part of my role was like, you know, I had been through that process before, so I kind of like, you know, could reassure the them a little bit that, yeah, you can start a company. Yeah, you know, we can, we can pitch this. There's enough here. Um, just sort of basic stuff like that. And then just, you know, the starting a company, it's, it is a little bit mysterious, right? Like, what do you have to actually do? How, what do you have to say to VCs? That kind of thing. It's not rocket science, but it is, um, something where, you know, doing it once does help a little bit. So that was partially my role as well, is just to be, you know, someone to say, yeah, let's do it, let's go. Okay.

I think out of all the 4 co-founders, you had the most industry experience. Yeah, I think you did, right?

Oh yeah.

Which is pretty cool. Yeah, I think I did get fundraising to arms race. It's always kind of the words are changing. Yeah, the deck format, the expectations are always changing. So yeah, that would have worked 10 years ago probably don't work now.

But, uh, yeah, actually it was a lot— I think it was a lot easier. Um, just there was more money when Hexagon started than when, um, 23andMe started, especially for genomics, right? When 23andMe started, there were no big genomics companies really, at least that I can think of. I mean, Myriad was the really big one, right? $2 or $3 billion. That was like a huge diagnostics company, and everything else was dwarfed by Myriad Genetics. So nobody really thought there was a lot of money to be made there. And there were no DTC health companies really, these big ones. But when Hexagon started, or today, you can point to like a lot of genomics companies. You can probably, I mean, Moderna, I guess, is a good example of something like that. But it's really a different world now. So yeah, it was, it was a different experience and a lot more VCs as well.

Yeah, I think, I mean, mid-2000s, you had this whole sequencing wars. So nobody even knew what platform is going to work out. And then I think Illumina started winning around 2012, maybe. I don't know. I don't know.

I think it was before that.

Before that. You know better than me. And then that kind of then enabled all these other companies to build on top of Illumina, like Hexagon. And so I think Proteomics Proteomics is right now, it's in a platform war. We'll see which one wins. Whoever wins that proteomics platform war, hopefully it's— we're an investor in Glyphic. So hopefully it's Glyphic. Oh, cool. So hopefully Glyphic wins. But we'll see. And then that will probably enable a whole new generation of companies. But you know, for Hexagon, maybe we could use Hexagon and also 23andMe as kind of case studies for understanding the role of computation in biology. And so what was the premise of Hexagon, uh, when it got started?

So yeah, the premise was that we could do, um, well, first of all, we should, um, sequence the Earth because contained within these fungi, right, are a lot of valuable drugs, but we've only sequenced 0.1% of the fungi in the world, and we've only seen, I don't know what percentage, but a fraction of a percent of the molecules that they make. But within that, we've seen penicillin, the statins, cyclosporine, like some really important drugs have come out of those. So the idea was like, if you could examine the genomes of all of these fungi, And then you could choose the most interesting drugs instead of just sort of finding them by chance. Sort of penicillin is obviously the most famous example of that. But the other natural products that you would have heard of, mostly they're found by chance. You know, you just sort of grind something up and see if it has a biological effect. And that's how the statins were discovered, for example. So it's sort of like about systematizing that process. Scaling up, using computation and automation to make something that was more like a reproducible process for uncovering drugs from the, from the global metagenome. And I, I don't know, I don't know that everyone likes this analogy, but for me, I really like the analogy of building a factory because I think the good thing about a factory is that you can measure the inputs and the outputs. And you can make it reproducible. And I feel like there are a lot of advantages to that compared to a process where you're sort of relying on some inspiration or some discovery to make the next drug. I think I would prefer to just scale up a process and improve all of the steps of the process and make it work, um, you know, make it work reproducibly so we can find a lot of drugs and not just, you know, get lucky and find one drug. Yep.

I think Mark DiPristo has a good analogy. He sees it as like a Ferrari team. Interesting. It's actually a great analogy. And like every week he wants to run his platform, a big hat, and see how fast you can go like a Ferrari car. And, you know, Enzo Ferrari though was brutal. He's pretty brutal though. So hopefully you balance that out. But yeah, I totally agree on the concept of industrializing biology and then exploring the full space, especially for microbes, where is this something like 99% of microbes can't be cultured under standard procedures or 96, some big number.

People say that, but I'm not sure if that's true, actually. Yeah, there's some controversy about that number, actually. So I'm not sure what the answer is, but certainly many cannot be.

It's a big number that can't be cultured. Through electrical means. So that's why sequencing's so valuable. Yeah. Just sequence everything. And then now the hard part, this is what Hexagon solves, expressing those biosynthetic clusters in your platform. And then you can express, you can mix and match and you get to express any natural product you want.

Yeah.

And so maybe you can also talk about the role of data in that whole process. You sequence the world. At least the microbial world. And then you have to then find what's valuable in this milieu of DNA and then express it and then try to make a drug out of it. Can you talk about that whole process from a platform perspective and also a team perspective? Because the one thing I had to brag about Hexagon, I think it might have the best software engineering team in the world in biotech.

Probably. Yeah.

I think it does. Brian, great job. I seriously, Hexagon has the best software engineers in biotech, and it's really crazy how you hire these people and retain them. So maybe you could also talk about that too, that building the data platform out, but also building a team to build that platform out.

Yeah. Yeah. I think, yeah, you gotta sort of differentiate between the software problems and the computational biology problems. I think some biotechs will sort of merge them. So they'll have sort of bioinformatics in charge of like their databases, but also like computing over their databases. I think we've been pretty strict about separating software and computational biology because I think frankly computational biologists are not usually good at software and software engineers shouldn't be messing with biology probably without training. So I think there is, um, it is good to divide them up and get specialists for each if you really want to scale it up. And I think, yeah, we've, we hired 4 of these engineers from Palantir, very senior, among the most senior engineers at Palantir actually, who had left sort of after, you know, many, many years sort of building up Palantir. And I think part of the reason they joined is just they were excited about sequencing the Earth and just the concept of finding a microbe, decoding its genome, looking for a drug and expressing it in another organism. There's something appealing about that. And I think, you know, it's, it's a little bit Jurassic Park, right? I think it's interesting that Ginkgo talks about Jurassic Park so much because I think it's just a cool— it's just a concept that grabs a lot of people, this Jurassic Park concept of making biology. So I guess synthetic biology really is what it, what it boils down to. So yeah, we brought them on. Um, we also gave them honestly a lot of responsibility, um, and sort of took them on almost as another founder. So they came as a group of 4, so I won't say like one of them is a founder, but as a group they, they came on at sort of like founder-level responsibility for building out the infrastructure, the LIMS, the just the data architecture. All of that stuff was, was on their, on their shoulders. And that was great for me because then my team could focus on computational biology, right, which is what, what I'm really trained to do. I'm not, not trained to do software. Um, I don't know how to be a software engineer, and nor do I want to be. So that, that really helps with, with hiring for, for my team. Interesting.

So you guys hired like a software Borg? That's the lesson, just Hire a Borg of software engineers if you can find them. That's the lesson for Hexagon, I'll say, in team building.

But yeah, and I think honestly, like, a lot of companies could not have hired these guys because we had to bring them on as senior people in the company. Most biotechs will hire, you know, according to their pay scale or whatever. It's okay, you're an engineer, you are worth whatever percent. That's how, you know, this must be laid out. But we just decided these guys were transformative and we wanted them and we were willing to sort of sacrifice our— I don't know, sacrifice is even the wrong word, right? It's like we thought they would substantially increase the probability that the companies would succeed. So then it's very logical to give them, you know, um, you know, their fair share of the company, right?

Yeah, but I think Hexagon was one of the first movers to like just, you know, pay for talent on the software side. And you're seeing even companies like ICON now led by Roger Perlmutter, and they're bragging about how, man, we have all these software engineers. And it's like, this is Roger Perlmutter. He's like old school biotech bragging about his software engineers he has on his company. So I think Hexagon definitely, uh, another pioneer in terms of like how to build a platform in biotech.

And so maybe, yeah, someone must have had words with Roger Perlmutter because I think that was not his attitude.

Somebody, somebody educated somebody. I don't know. He retired and then he must have met, he, he must have talked to somebody and somebody must have opened his eyes.

Uh, yeah. And now EQRX too, right?

Similar.

It's a, Lux is a little more, he's a little more open-minded and flexible. Also Jacob Oppenheim is there. Actually, we're gonna have, we're gonna have Jacob Oppenheim here in a week to talk about Ecorex. He's incredible. He's incredible. Yeah, he's great. Uh, but, uh, uh, you introduced me actually. So I, I've kept in touch. I kept in touch with Jacob. Jacob's one of these superstars too. I think he manages like a team of 30, 40 people now at Ecorex. It's crazy. It's crazy actually. Uh, but maybe you could talk about the value of data, the value of software. In drug development, at least in the context of Hexagon? Nothing confidential, but maybe just a high level, like what has it enabled you that maybe another company couldn't have done, you know, 10 years ago?

Yeah, I think, honestly, I think there are sort of two ways to do a biotech. And probably you want to do one or the other, not like halfway. I think one way is to just run it like a lab. And then you can just use whatever software you want. And just essentially have— use the brains of your scientists to drive programs forward. You know, let's say you're starting with an asset or you're starting with an idea and you want to, um, do medicinal chemistry and improve on this asset. Then I don't know if you want to hire, you know, expensive software engineers. And if you hire sort of some software engineers, I I don't know if it really is going to move the needle for you, right? So I think there is maybe a distinction between the type of biotech that needs software and the type that doesn't. And I think it's definitely possible that biotechs will start hiring software engineers because they feel like that's a thing that biotechs need to do now, but not really reap the benefits. Because I think honestly, you only reap the benefits when you get to scale. Until you get to scale. You sort of, you're sort of adding layers of abstraction or layers of indirection on top of the work. So if your scientist wants to do a new type of experiment, now they have to sort of have a new table or something in the database. They have to somehow talk to the software engineers to get, you know, to get us set up so that they can do that experiment. So you're slowing, you could be slowing yourself down. With Hexagon, right, we wanted to build this pipeline that we, we pretty much knew what the pipeline was from the beginning. You know, you put genomes in one end and you get drugs out the other end, and there are, you know, 10 steps in between that are not totally fixed, but we wanted to do it hundreds, thousands of times. We wanted to repeat the process. So for companies like that where it's like a true platform and you want to repeat something, um, that's where it becomes really important to have good software and good algorithms, honestly, good data analysis on top of things. But for some companies, I, yeah, I'm not sure. I feel like there's a tendency to sort of recommend it for, you know, data is necessary for every biotech. I guess I'm not sure if that's really the case. I think you have to make sure you know what you're, what you want before you start on that road. As a quick example, I won't name the company just in case, but somebody had a at a big company told me that they had a lot of— they had spent a lot on software, like a ton of people, tens of people, a good 20% of their company was, was software and data science. And they really weren't getting a lot out of it because they had too many projects. So then the number of software engineers per project was so low that nobody was really satisfied with the software. So even though it was costing them a ton of money, they felt like it was— they probably should have run it as, you know, 10 independent labs, almost each one responsible for its own business. And that would have been freeing, I think. So anyway, that's a hard question.

Yeah, I think it's just like a lot of these— if they call them AI drug discovery companies, they kind of look like CROs in the beginning. Yeah, you know, they kind of, you know, you know, in situ, Hexagon, they kind of in they're kind of spending all this money, recursion, all this money to build all these fancy tools. And it's like, oh, is this going to be valuable one day? I hope so. And that's the whole premise at scale. It becomes valuable. And so do you have a framework in terms of thinking about, you've alluded to this, is like some biotech companies have like a hypothesis, say we have an antibody that binds tau. Okay, cool. Let's go do that. Boom. We don't need data for that. We just got to go to the trials. Whereas there's another premise of saying, let's sequence the world and we can find all these cool, you can find all this new chemical matter. Do you have a framework just for founders on how to match the platform to the problem? Yeah.

I think you kind of know what, what you need, or you should know what you need. I think, you know, Big Hat versus Denali, I think, is a good sort of, um, comparison. So both very smart companies, both working on antibodies, but I think really, really different approaches. And I don't know that one is right and one is wrong. Certainly the Big Hat model is the one that I would be drawn to, right? They want to build a system for, you know, making better antibodies quickly improving the process, improving the tools. And my sort of— I don't have insider information about Tonalii, but my— what I've heard is that it's sort of run more like academic labs, and they don't really have centralized infrastructure, data, or whatever. But I think, you know, it's possible that this is actually the right way for Tonalii to run their— to run things. They have some extremely talented scientists there. And they know how to make antibodies and that's what they're doing. They're not trying to build a process, at least I don't, I don't think so from the outside. So I think you have to kind of like work with your strengths a little bit, right? Like Mark dePristo's and Peyton Greensides, their strength is clearly like computation and thinking about things in that systematic way. Um, and the Denali people, their strength is, you know, they know antibodies inside and out, and they know Alzheimer's and Parkinson's inside and out, and those, those neuro diseases. So yeah, I think as long as you work into your strengths, you— it's okay. Um, you just don't want to like hire software engineers if you don't know what a software engineer is for, because then you end up with a big software engineering team that you don't know— that you can't derive value from, you know.

Totally agree. I think that kind of comes down to the business model then. Maybe we could Gears to the next section. We're like, we're an investor in Big Hat. Thanks for introducing to Mark too. I introduced you to Mark. Wow. You did. You introduced me to Mark. Before we started Big Hat, it was like he was at Google. Yeah. And he was just trying to fish for ideas. And so we actually talked a lot about enzymes. He was, as an early idea, but he ended up doing antibodies and he found his perfect match co-founder of Payant. Payant is a perfect match for Mark. Work. Yeah, seriously. But maybe we could talk about business models of biotech and AI. Yeah. And, you know, maybe Denali is very focused on just therapeutics in the clinic, and that's one business model. And then Big Hat's building a platform out that can then get partnerships, get scale, and hopefully, you know, one day build up a larger pipeline of medicines. Yeah. How do you think about business models in drug development and maybe business models in general? Because you've been part of a lot of creative companies over—

Yeah, I think there's no real wrong way to do it. I think traditionally, sort of partnerships with pharma were tough because they just took a long time, like pharma was slow. And they wouldn't really pay for early assets, like certainly wouldn't pay for data. They would pay for molecules, right? Pay for actual assets. That's been sort of the way it's been for a long time. I think it's changing a little bit. I don't think it's as bad now as it used to be. But then I also think there's a lot more money available, right? As everyone knows, there's just a ton of money floating around. So if you do want to build your own program, your own internal program, you can get enough money to do that, I think, these days. Whereas for, for example, with 23andMe, it would have been really difficult for us to raise enough money to become, to go at therapeutics really hard. And in the first, I don't know, 5, 8 years, because the money, I don't think the money wasn't really there, I don't think. But now I think it's easier. So I don't know, kind of a wishy-washy answer, but if the deal is good, you know, it's okay to partner. There's no— there's nothing wrong with partnering, right?

Yep. I think there's always historically of drug development, there's always ebbs and flows. And yeah, I think there's a project— I used to watch Project Runway. So like Heidi Klum, we see what one— one day you're hot, one day you're not, or something. I like Heidi Klum, but she would say like she had that quote or something. And so sometimes in biotech, CRISPR is really hot. I'm assuming CRISPR is not going to be so hot in 10 years or one of these years. Gene therapies is in a dark age and the in-druggable genome, right? It's like crazy. 10 years ago, drug Keras, that's ridiculous. Now you have Amgen and Merati. And so at least for Hexagon, maybe for your business model, you have such a wide-ranging platform. You sequence the world, you get all this chemical matter. And what, how did you focus on a specific problem? You could do infectious disease, you could do autoimmunity, you could do oncology. You could do, you can go after KRAS, you can go after SHP2, you can go after a whole host of different targets. How did you get that process of honing that down over the years?

Yeah, so it's kind of simple in a way. Like, um, anti-infectives are good for hexagon because we make these very potent natural products. And sort of traditionally, the strength of natural products is they are very, very specific for one enzyme, and they are often just manufactured as chemical weapons and You know, sort of famously, the most toxic molecules in the world are natural products like botulinum and ricin and these things. So it's a really good fit. I think two-thirds of anti-infectives are actually natural products still. So as long as it's very specific to the bacterium or the fungus, you, you can make a great anti-infective that way. I think it's kind of like almost the the natural way to do it. And then oncology is pretty similar, right? You want to find, um, something undruggable ideally, and then match it to a natural product that drugs it. And there's a history of natural products interrupting protein-protein interactions and hitting undruggable targets and just generally disobeying the laws of, of drug development, you know, being very large but still getting across the cell membrane. Instead of directly interrupting a protein-protein interaction, they can grab another protein from within the cytosol and use it to interrupt the protein-protein interaction. Things that you really can't design with, you know, traditional chemical synthesis methods, like you can't even use AI to get there. You have to use sort of hundreds of millions of years of evolution to get there. So that's really where the strength of the platform is, is in hitting getting these undruggable targets and protein-protein interactions and just sort of these new scaffolds that you can never really develop in the traditional way. We can find them in nature and then leverage the work that evolution has done to make those drugs.

I think experiments of nature are very powerful, whether you're trying to discover new gene editing technologies or new cancer drugs, letting nature invent it is a good starting point. You can optimize it later on. And really awesome. I think we have a clubhouse. Hey, Dikesh, if you want to join the stage, you're welcome. But I think we have a question. So let's see. Let's see how this question goes. But sorry for— Hey, Roxana, how are you? If you have a question for Brian, love to hear it.

Yeah, thanks a lot. So yeah, so as a PhD student in genomics and also part of an institute that does a lot of like sequencing and data and stuff like that, some of the things that I find hardest when thinking about companies and startups from this area is actually is how do you get from that amount of data to an actual product? And I think what you did that is great is you combine this with wet lab. So I was just wondering if I I looked at your co-founder and she also seems to have a computational genomics background. So I was just wondering, how did you go from, from that to integrating wet lab? And yeah, because like you, yeah, how did you, how did you integrate that?

Yeah, that's, that's really important points. I think there are companies that can sort of be almost entirely dry, right? Like Atomwise and Schrödinger are true computational chemistry companies, but almost always you, you will want to partner with someone with wet lab experience. So Maureen, the CEO, had worked in wet and dry labs. So she had worked, I think, most recently in Ron Davis' lab at the Genome Center. And then Colin Harvey, our other co-founder, is a chemist who was working with Maureen. So he was totally wet lab for all of his, his training. But I think it's a really good point that you can't just do genomics and have a company. You have to have some kind of product in mind. So if you think about like all the great genomics companies, almost all of them are in diagnostics because For companies like Color, Invitae, Counsul, if you can sort of master the software and maybe the automation and the variant analysis part of it, then you can sell that, right? You're done. You don't need a very strong wet lab component to make that work. So it's really, it's kind of partially a question if those diagnostics companies are if that's covered, right? If we've covered diagnostics already and there's a lot more free space in therapeutics, then you definitely need some wet lab support.

You can't just do it with dry lab. Okay, thanks a lot. I wasn't really aware of your co-founder who was a chemist, but, uh, did you, did you have the platform sort of already invented or patented before starting the company, or did you kind of develop it after kind of incorporating or something?

Yeah. So the heterologous expression platform was working at Stanford and kind of like a prototype stage. You know, it wasn't automated or really industrialized at all, but it got you from beginning to end a few times in Colin's hands. So we're sort of confident that we could extend that work and scale it up. But yeah, there is a big advantage to doing that kind of work within a university. So the university ends up owning part of the company, obviously, because they own the inventions that happen at Stanford or wherever. But honestly, I think in most cases that's probably worth it because you get to spend, you know, grant money developing ideas and you don't have to, you know, answer to investors every year about why you're not making progress. You get to do a little bit more exploration instead of just picking something and going for it. So yeah, I think that's actually a good way to do it.

Yeah, thanks a lot, Brian.

That's very awesome. Great question. Um, I think we're coming on the hour. Yeah. So Brian, do you have any like kind of final thoughts or advice from your experience? I think some useful tidbits are like how to set culture. I think, um, you know, yeah, How to build a board, right? Build— that may be something else to talk about. But any kind of final, like, tidbits of advice for, you know, potential founders out there?

Um, I would say, uh, you know, choose your first few employees very carefully. It's really the most important thing. Um, and I think over the years, like, probably when— I don't know, not that long ago, right? It wouldn't be crazy for founder of a company, especially a software company, to have, I don't know, 10% of the company, and then the first employees to have point something. So some very low percentage, 0.1, 0.3%. I think things are changing there. And I think the companies that sort of recognize that will, will do better because I think there's a lot of demand for talent now. And it's a lot clearer. Who has talent and just who's in demand. So you have to be willing to, you know, give your first employees enough equity that they feel like they're part of the team, you know, have responsibility. And then you get better people and that pays back, pays itself back, you know. I feel like a lot of companies make that mistake. They just don't pay for talent, I think.

Yeah, but I think that highlights your career, being part of all these great talent hubs from Trinity to Stanford, 23andMe, Hexagon, and probably beyond. I think you find a good way to be around talented people. Seriously, seriously, you do a very good job. I think one thing I'm observing Totally agree on the talent side hiring. I'm also seeing companies like get rid of vesting. Yeah, like modify, modify vesting actually. Say, hey, I like that. Screw this vesting thing. Just here's, here's half your shares now, half your shares in a year, and let's not do this 4-year vesting cliff thing. And so I think, I think that—

interesting. I've never heard of that.

That's interesting. There's weird, uh, some— it's just we're all in biotech at least. It's a apocalypse for talent. It's really hard. Yeah, it's really— it's hard to hire RAs, to be honest.

And so, uh, well, it is and it isn't, right? If you paid 25% above market, it wouldn't be hard. That's the thing, right? It's like you can get the best RA in the world if you— and it's not that expensive.

It's just, yeah, but then you gotta pay for talent, but then I'm poaching it from you, or I'm poaching it from, you know, a friend's company, so it becomes kind of a It's a little bit of zero, I think. SEO, it's a whole other conversation. But yeah, maybe, maybe it's really— hiring is really hard though right now. It's like the hardest thing, hardest thing about a startup, build a startup right now is hiring. But maybe to wrap it all up, for Hexagon, or maybe just personally, anything you're looking forward to over the next year?

Oh man. I don't know. I'm looking forward to, I don't know, developing the platform and I don't know. I got, I got no answer to that one. Cool. Maybe we're, we're just in the middle of it.

We're in the thick of it. I'm assuming it might be some of the confidential stuff, but I'm assuming Hexagons are going to have some big announcements over the next year. So, yeah, absolutely. Really looking forward to it.

Yeah, it's interesting. You think you're, you know, you think you've done your inflection and there's always another inflection point. There's always another way to double something.

And it's just, yeah, there's always a bigger boss. It's hard to look too far ahead. Yeah, exactly. There's always a bigger— when you play a video game, there's always a bigger boss.

There's a bigger boss. Exactly.

Just another quarter. But cool. This is really fun, Brian. We should all— Nikesh, you and I should all get lunch in Palo Alto one of these days.

That'd be cool. Sounds great.

If you're in Berkeley or SF, just give me the ping on it. But, uh, hope people enjoyed this and found this useful. And yeah, Brian, thanks for doing this, man, and we'll talk soon.

Appreciate it. All right, thanks a lot, Josh. Talk to you later.

Well, have a great day, everyone.

Bye.