Tough Tech Today with Meyen and Miller
Tough Tech Today with Meyen and Miller
Creating new and better medicines, featuring Virginia Burger of New Equilibrium Biosciences
In one sentence...
Overview
Show Notes
- Episode page, transcript, and podcast listening links: https://toughtechtoday.com/creating-new-medicines/
- Virginia Burger on LinkedIn: https://www.linkedin.com/in/virginiaburger
- New Equilibrium Biosciences (company): https://newequilibriumbio.com
- Watch this show on Youtube: https://youtu.be/PUa9f8UORvc
Credit Roll
- Producers: Jonathan "JMill" Miller and Forrest
- Guest: Virginia Burger
- Hosts: JMill and Forrest Meyen
- Contributing Expert: Malvika V. Miller
- Editing: Giro Studio and JMill
- Transcript: Alan Yan and JMill
- Blog Author: JMill
- Art Design: JMill
Topic Timecodes
[21:16] Training data for AI models
So gaming to proteins!
Virginia Burger:I went to a talk by a professor who's talking about the protein folding problem. And then I was like, "Oh, this is what I want to work on." And then he's like, "Oh, you've been doing all these programming courses. That's convenient." And so he said, "I think the field is called Computational Biology."
Announcer:Welcome to Tough Tech Today with Meyen and Miller. This is the premier show featuring trailblazers who are building technologies today to solve tomorrow's toughest challenges.
Jonathan:Our guest is Dr. Virginia Burger, co-founder and CEO of New Equilibrium Biosciences. And she's going to take us on a journey into creating new medicines with computational biology. This episode was recorded in 2021, and is a time capsule of where Dr. Burger was then having raised a $10 million series A for her company. In the time since recording, New Equilibrium has grown in headcount and and resources, including deploying a high performance computing cluster for simulating intrinsically disordered proteins. We'll get into what those are soon. We are joined by our guest expert for biotech, Dr. Malvika Miller. Let's dive into the life sciences.
Malvika Miller:We're really excited to have you Virginia to chat with us. And of course, thank you, Jonathan and Forrest for having me on board, we'd love to kick off and ask Virginia to introduce yourself, and tell us what you're working on what your priorities are. And we'll go from there.
Virginia Burger:Thank you all for having me, it's a real treat to get to be here and get to talk about what we're working on, because I'm so excited about it. So I am the co-founder, as you said of New Equilibrium, where our mission is to discover new medicines for life-threatening diseases, and really help some patient groups with large unmet needs. And we're working on this class of proteins called intrinsically disordered proteins, which are unique in that they don't have a folded structure. And most proteins until the 80s or so were really thought to always be one single folded structure. And there was this lock-and-key hypothesis. And then, in the last 20 to 30 years, people have realized like, oh, proteins take on all these different shapes. And those are called intrinsically disordered proteins. And they're also very involved in disease. And we can get into the science when you guys start asking me questions, but they're involved in so many different diseases, and people aren't able to target them because they break this paradigm of proteins supposed to have a one folded structure. And so this has led to all these different challenges and dragging them and we add New Equilibrium are applying a computational experimental approach to actually see all the different confirmations and use this to guide drug design. And I come from computational biology. And that's actually how I got interested in intrinsically disordered proteins. It was just the first thing that was mentioned to me when I started graduate school. And I was like, "Oh, those are cool. I'll work on them." And then I just kept working on them.
Forrest Meyen:So what percentage of these proteins are intrinsically disordered? Like, is it just a tiny fraction of them? Or?
Virginia Burger:No, it depends how you count, but it's probably at least a third in humans, so it's not a small deal. And a lot of proteins will have a long, disordered region, which has at least 30 amino acids that are connected and don't have a structure. And this is really, really prevalent in biology. And so, so many of the different targets we're looking at have intrinsically disordered regions that people would love to target, but can't. So they're really common. Something like 60% of proteins that are involved in diabetes or different other diseases tend to be intrinsically started. I'm forgetting what number corresponds to what disease but it's all way above 50% that would be good targets.
Jonathan:So then it correlates that intrinsically disordered proteins are, would we label them as bad? Or that they are just there different?
Malvika Miller:Yeah we discussed, maybe we can ask Virginia, what is intrinsically disordered even mean? Like taking a step back?
Virginia Burger:Yeah. So yeah, they're not bad, they're good. But when they misfunction, that's bad, just like when any other protein misfunctions, but with intrinsically disordered proteins, we don't know how to solve it when they do. But they're definitely really good proteins. So all it really means is that they don't have one folded structure. And the hypothesis has always been like, proteins have a single amino acid sequence which nobody's gonna argue with that. But that that folds into one structure, and that's been the protein folding problem that people have been working for the last 100 years of how do we take a protein sequence then we predict structure, but it turns out only about maybe two-thirds of human proteins do have a folded structure and the others, you have a sequence, and then it predicts what corresponds to many, many different confirmations instead of one structure. And all of these confirmations can have different functions. And the technical definition that I use is that a folded protein has a funnel like free energy landscape with one folded protein at the bottom of the funnel. And intrinsically disordered proteins have flat free energy landscapes with lots of different confirmations that are easily accessible. And our logo is the free energy landscape of an intrinsically disordered protein.
Jonathan:But tell us... maybe elaborate now, essentially, you mentioned the logo of New Equilibrium. And you describe some really interesting and deep science. How does one go about commercializing this, because that's what I mean you're running a startup. So it's inherently a commercialization attempt, and also a successful fundraise that recently occurred. So can you tell us more about how you can make the science go from the lab to the market?
Virginia Burger:Yeah, so we're commercializing it, because these proteins are involved in so many diseases. And what we're making is a platform that allows us to design drugs for these proteins. And so within that platform, we can both have internal drug discovery products, and just be basically a pharmaceutical company, or we can also partner and have other companies use some of the compounds that have come out of our platform. But there's a lot of commercial opportunities, because they're just so highly involved in so many diseases. So many really famous proteins, like p53, is intrinsically disordered, Myc is intrinsically disordered. K-Ras has a really small region of disorder that's very functional at the end. So like all of these super hot cancer targets, would be more easily targetable, if New Equilibrium actually were to be successful. So that's what we're trying to do is be successful and help.
Jonathan:You see, then you have this, this who's who list of the sort of... I'd say sort of the villains, right? You listed the names of them. And so then with New Equilibrium, the goal is to... to use an interesting... there's a couple of Buzz phrases, I'd be glad to sort of demystify those of computational chemistry, quantum computing, artificial intelligence, to be able to identify and sort of like "zap" these villains, these proteins that we need to figure out how to handle.
Virginia Burger:I'd say not identify—we know them—but they're experimentally invisible. So we want to uncover them. And then zap them in different ways. But yeah, definitely. There is a who's who listed proteins. And so that's made, I think, part of becoming a company has made it easier that we already know that intrinsically to certain proteins are very important. My co-founder has been working on them since 1999, and really building up the field around why are these proteins interesting? The first question was, do these proteins even exist, and back in the early 50s and 60s, when people were getting the first crystal structures, there would sometimes be intrinsically disordered regions on these proteins, and they'd say, "Oh, this isn't letting us form a crystal. So we can't solve its structure." And they would just cut it off and say it was an artifact. And so that was just kind of going on and on. And then eventually, I guess, in the 80s and 90s, people were seeing them more and more and realizing like, "oh, maybe these actually play functional roles." And so that turned into the field of intrinsically disordered proteins. And it's gotten to a few name changes, like when I first started studying them, in beginning grad school, people called them natively unfolded, and now it's intrinsically disordered. And so I think with any field, it kind of takes a while for it to get a name. And then once the name settles, it becomes a real field. And, and we're there now, because we're seeing intrinsically disordered proteins all the time. We're not gonna let the name change.
Jonathan:But it's important to have sort of a name that ties it together, I think, in advance of the field. And I understand your co-founder, sort of literally wrote the book on some of this. Is that a correct assessment?
Virginia Burger:Yeah, yeah, he wrote a book. It's behind me. Called "Structure and Function of Intrinsically Disordered Proteins." And so like he kind of... it was, I think, around 2008, that he... I should know what year he wrote this. Anyways, 2010, he wrote this. And it kind of brought together so much of what had happened in the field until then, and it's really important because it like brings in the function that like all of these different confirmations have different functions, and they allow the protein to function. And the most important thing is that they're often involved in signaling. And so by being one protein that can take on many different shapes, they're able to bind to many different partners. And the most famous intrinsically disordered protein is p53. And it has hundreds of different partners that it binds to that are other proteins, like DNA and RNA. And so when it's binding to these different things, it's passing on different signals. And so if there's any mutation in p53 that causes it to start passing on signals more frequently, the whole cell will start doing the wrong things, because it has the wrong signaling. And that's why they get so involved in disease. Um, so a lot of signaling proteins are disordered. But then there's like little fun stories about them to where they can unravel and pass through a channel. And sometimes like they unravel, they reach their channel, and then they move things around on the other cell that they've been able to reach into and change things and then come back. And so this disorder really opens up a lot of possibilities. I mean, just like humans, like we're better off, you know, being able to move around than not usually(not to, you know... it's horrible when we can move around). So proteins like to do the same thing. Yeah, so he wrote the book and then when we had our dinner... we met, and I was like, will you be my co-founder? And then he emailied me back and said yes. But then he was in Brussels. And so I went back to the US. And then I was like, "Well, I guess I'll come back to Brussels." And so like, two or three weeks later, I went back, and then we went out for dinner. And then he gave me the book, and he inscribed it and it was really kind of like a special moment to get that.
Jonathan:I like how you almost blush when you say like, "can you be my co-founder?" Like establishing that kind relationship is not for the faint of heart. I mean, that's quite a commitment to be putting reputations, time, and everything on the line to advance this idea. And this was what circa 2018, 2019?
Virginia Burger:It was 2019 and August or July. Yeah it was July, and yeah, the thing that was the most meaningful me to was the reputation because he kind of like... I mean I didn't actually expect him to put in nearly as much time as he's done. But having him say, like,"Yeah, I actually believe in what you're doing. I think it's the way to go based on my 20 years of working on this" was really cool. And then he does work so much... like we talk every single day, like we got each other's WhatsApp. And then it just went from there: that it's really worked out to be like way more than I would have expected and really great to have happen.
Malvika Miller:Virginia, I have a question for you related to what you were kind of mentioning: you're bringing your computational background to a problem where I understand like p53, Myc, these are undruggable proteins and changing confirmations all the time or understanding that and you know, that's a lot of structural biology. Can you tell us a little bit about the computational side a little bit more, because, I mean, there's just so many buzzwords, like Jonathan mentioned being thrown out—quantum computing, AI—like, what are those techniques bringing to this field, versus like doing like high throughput screening, or sort of trying to do more classical ways to undrug these undruggable proteins that I understand are obviously, super hot and exciting targets, but would love to get your thinking on this computational wave here?
Virginia Burger:Yeah. So typically, what people do when they want to simulate a protein and see how it moves around is they take a starting structure, and then there's a function called force field. And that describes the force on every atom in the system as your protein and the solution around it. And then you use that force field and you calculate the force, and then everything moves forward one step because of force equals mass times acceleration, so you can divide out and get the new position. It involves integration but that...
Forrest Meyen:Just like high school physics, right?
Jonathan:That's awesome.
Virginia Burger:Yeah, it's Newton's laws. So you move it forward using this formula, and then you go to the next step, and you go to the next step, and you do this. The step length is usually like two femtoseconds. So and that's a fast time step. Some people like to go smaller.
Forrest Meyen:What's a femtosecond?
Virginia Burger:It's 10 to the negative 15 seconds. So you're moving really, really slowly. But you have to move really slowly, because the that all of the bonds are vibrating. And if you move faster than the bonds are vibrating, then the whole simulation gets messed up, just like when you're watching one of those movies where people have the flip frames, like if you skip over 10 frames, who knows what's gonna happen. So I guess if you don't skip over 10 frames, we have the person take one step, and then all of a sudden they're running, it just doesn't work. So people move at this really small scale, then you have to get at least a microsecond, which is 10 to the negative six. And so to go from 10 to the negative nine to 10 to the negative six, it's a lot of simulation steps. And so that's how people have simulated proteins moving. And the key part is one, doing broad sampling, and two, having an accurate forcefield. And that's where New Equilibrium is coming in with all of our buzzwords, because traditionally, these force fields were developed in the 70s, and actually, Martin Carr plusa couple of other people got the Nobel Prize in 2013, or 14, for the work that they've done in developing this molecular dynamic simulation architecture in the 70s. And those are functions that are based on the physics of proteins, it looks at how bonds move, and how bonds stretch and how bonds twist. So stretching, twisting, and bending. And then also non bonded interactions that have to do with electrostatics, and everything. And so this goes into this one function that you calculate in every single step. And I've gotten really detailed on the computations. But basically, you do these simulations, and then you pull out if it's an intrinsically disordered protein, all of the confirmations that are sampled a lot, and you say like, "Okay, this protein samples, like these 20 confirmations," and then you can try to show that it agrees with whatever experimental data is available. The challenge has been that while this works really well for folding proteins, the forcefield, this big function, is based on folded proteins. All of the training data that went into it came from folded proteins. And so over the years, people have found that when they simulate intrinsically disordered proteins, they're getting too much structure, everything's more compact than it should be according to experiment. And everything has more basic protein secondary structures, like Ulysses and beta sheets. And so they don't match what's in the lab, because there's this huge bias of the computations to folded proteins, and so when I first started thinking about New Equilibrium, I was reading a lot about like startups and what startups should do and how to have an idea, because I knew we needed a better way to simulate intrinsically disordered proteins. But I didn't know how. And then I read this, like blog or Medium post by Elon Musk about first principles thinking. And the idea was that you like, you take the problem. And then you break it down. And you're like, Well, what tools do I have in 2018, or 19, to solve this problem instead of iterating? On how people had solved it previously, because there has been a lot of iteration where people have said, like, okay, these proteins don't work well or these simulations didn't work well for intrinsically disordered proteins, let's change this parameter or run it this way. But it's always been kind of like gradually tweaking on what was already there. And so I was like, "Oh, well, Elon Musk knows how to build a startup. So I'll do what he does, like, what do we have in 2019? That is a tool that I should be using if we're going to be building a new way to do things." And AI was just the buzzword everywhere. And so then I started thinking,"Okay, well, how do we use AI for simulations?" And that work had been going on in the small molecule field of simulating energies for small molecules. And so I was like, "Okay, well, how do we expand that to proteins, that's going to be what New Equilibrium does." And so that's
where the AI came in:to get the accurate agreement and to move away from training data on folded proteins, but to look at training data on quantum chemistry and other things that we can get now better than we could in the 70s for intrinsically disordered proteins. And so that's a big part of it. And then the quantum computing kind of thinking in the future that we're also going to really need to accelerate these, and so when we're making our decisions now about these new algorithms we're building we should be thinking about like, "well, will this actually work on a quantum computer? What what could we do differently so that it would" and so it's really... I think I said in the press release, it's building for the future, but that's what we're doing because we don't want to build something now when we know there's this great new technology coming and it's not even going to work for that. So we're keeping that in mind in everything we do.
Jonathan:On the way of thinking and building for the future... for the quantum computing applications. How do you think about what... is it like we're looking for opportunities where we can do a lot of factoring where that would be a great speed up? And then we know that that factoring works well, like Shur's algorithm on quantum computers. Is it thinking in terms of the algorithm design?
Virginia Burger:Yeah, thinking of the algorithm design. I guess... when I was programming, in grad school, we were using a lot of MATLAB, and MATLAB does really well with vectors. And so I'd taken some code I'd written in a different language and I was like we're just going to do everything here with vectors because we can, and it just totally sped up the code. And so what we're doing right now is we're actually hiring a person with a passion for quantum computing and having them spend four days a week doing their normal work. But then one day a week really saying like, "okay, this is how quantum computing is going to work. This is what I think would work," and really kind of guiding us in what would we be missing out on, if we weren't keeping that in mind? So we're pretty preliminary in there. But we're also just a preliminary company overall. So we're able to get that focus from the start.
Forrest Meyen:You mentioned that you need training data for developing your AI models. What sort of training data you actually input?
Virginia Burger:Yeah, so we use a quantum chemical data that's computed with different levels of accuracy on different types of peptides. And that's, as we saw, we built this very large database already. And as we're growing our algorithms, we really have to make sure that we might change our data at any point, because things change. So we're really trying to kind of grow the database as we grow our algorithms. So that we're really able to accurately model these proteins. And with AI, I think, if you have bad data, you're going to have bad algorithms. So like, I guess what I'm saying is I really like our quantum chemical data really, right now. And it's working well with the algorithms we have right now. But like, then I'll have a new idea. And then I think,"oh, we might need to change how we're making those data." But we definitely aren't caring, I guess, because it's kind of for IDPs, there is no experimental data that we could have changed from. So we had to build data in house. And then this gives us this freedom to really build the right data and build the models that match the data. Because a lot of times with AI, you already have a great data set, which is very, very useful. But in this case, we don't. So we're actually like kind of looking at the problem and then building the data and the algorithms together for the problem.
Malvika Miller:You're building your own training data in house.
Virginia Burger:Yeah.
Malvika Miller:That's great.
Virginia Burger:Yeah, we kind of had to.
Malvika Miller:Yeah, especially like you said, you're taking a problem that a lot of people have looked at before, but now in a completely new lens. How do you and your company think about partnerships with either obviously big pharma companies that are looking at targeting these undruggable targets or even tech companies, you have a lot of AI tools be it like Google or other other companies. So just just curious to know, your strategy on partnerships and how you think about that?
Virginia Burger:Yeah, so around actual pharmaceutical development, we want to partner as much as it will help us really accelerate our assets into the clinic. And if there's a team that is perfectly matched for a system that we're working on, we'd be very happy to discuss a partnership there. We're also very much working on internal development programs. And so our key focus is always those projects. And then around the technology, we're interested in partnering as much as it really makes sense that because with partnering, like we have our internal IP, the partners have what they're working on, and so we have to make sure that it's really a complement and something that we wouldn't want to be building ourselves. And with IDPs being kind of a new area, there's a lot that we're building ourselves. But there's also a lot of interesting technologies out there, and especially around different experimental strategies, because we're really building a lot of core expertise on the modeling. And then we're using existing experiments right now. And people will invent new methods for high-throughput screening against IDPs. And we're very interested in working with those types of people.
Forrest Meyen:So you just closed a funding round. Tell us about that. How much money did you raise? What did you raise it for?
Virginia Burger:Yes, so we raised $10 million. And we are using that for growing our team and increasing our current pipeline. So just bringing our current projects further, getting to more complicated experiments, and then pursuing partnerships, and just getting to understand really the needs of different pharmaceutical companies in this area more. But the really primary thing right now is growing the team. We're looking for both scientists—who are really hands on scientists—and we're also looking for people who are somewhat hands on, but really have a lot of experience in the field and have brought drugs to market before who could join us, and really bring in that expertise.
Forrest Meyen:How do you go about finding those people?
Virginia Burger:Yeah, I think one thing, just telling everybody that we're looking for them, doing podcasts, and...
Forrest Meyen:Alright, listeners, job opportunities. Where do they go? Your website?
Virginia Burger:Yes, they go to our website, or they just email. We're always interested in people and who especially have a background in either AI, quantum chemistry, or molecular dynamics simulations of proteins, but we want people who worked on all three areas, and maybe have expertise in one, but really have seen the whole problem before and can understand like how things connect, because we're very interdisciplinary, and the people who we've ended up hiring, have usually had some experience also doing experiments, or if they've been experimentalist, they've had some experience doing computation, because the team is so integrated, and the project is so integrated that we need to have people who really have demonstrated curiosity for the different areas of what we do.
Malvika Miller:So Virginia, you mentioned you're converging many different computational, but also experimental techniques. How do you stay on top of the field? Like, what are you doing? What are you reading to keep ahead?
Virginia Burger:That's, it's, it's hard. So I've been getting.... I don't know where this comes from but... Google Scholar sends you emails every day: I have seven topics that I get emails about every day. And that's very helpful. And then I read like Endpoints News, and the other things that send me mailing lists every day. And so that's part of it. But we are also starting a seminar series at New Equilibrium on intrinsically disordered protein simulations and drug discovery so that every month we'll have a talk about one of them from one of the leaders in the field, because I kind of felt like that was a way to bring it to us and also to build this community around those areas. Because like, in doing all of our kind of competitive landscaping, we've seen a lot of other companies in the space. And I don't think anybody's really directly competing. Like we could just learn more by sharing, because we're all working on slightly different areas. So we're trying to just really connect everybody in the IDP simulation and drug discovery areas. And so that'll help but yeah, lots of... like whenever I don't have meetings, I try to read papers, or at least titles of papers. Do you have advice for staying on top of things like other than that?
Malvika Miller:Me specifically? No, I think that sounds like you're doing what you need to do. And I think the seminar series sounds like a great idea. That's going to be internal to your company, and you'll invite some external speakers?
Virginia Burger:We're inviting external speakers, and we're opening it up to basically everybody but really the IDP community.
Malvika Miller:Okay, okay.
Virginia Burger:Yeah.
Malvika Miller:Yeah, it seems like a great way to figure out what is going on in the field and stay on top of it.
Virginia Burger:Yeah.
Malvika Miller:Jonathan, how do you feel about staying on top of things?
Jonathan:Yeah, well, I think the one you mentioned, Virginia, the Google Scholar hits and more broadly just Google search... you can set alerts... that is really helpful if you know what some of the terminology... that one's looking for it. Sounds like now that, Virginia, you and your team's field, is around so the inherently disordered proteins so you have consistent terminology which I think gonna make things a lot easier as the industry and the science rallies around that terminology. If you don't have that, that kind of like consistent way of looking for something then it's hit or miss but even in some of the groups that are on Clubhouse, etc, where there are opportunities to create sort of niche conversation areas: you know, ones like voice and live synchronous, and even some of the deep areas of Reddit: some of which are black holes, some of which are actually pretty nice communities, but those are areas too that one can seed a community if one doesn't yet exist. But yeah, well, I think with Endpoints at least for, generally, the biotech industry that rallies around that trade rag, so it must have been quite a fulfilling incentive set of emotions to see your work Virginia being highlighted in Endpoints.
Virginia Burger:Oh, yeah, I took a screenshot. Yeah, that was very exciting. Um, yeah, I think for intrinsically disordered proteins, the field, it's new, like it's been around since late 90s. And people have been working on it before, but it wasn't connected. And sometimes... not so much recently, but when I was a postdoc, I would read an article about something that wasn't intrinsically disordered protein but they never once mentioned that in the paper, and it would be from a different field like bacteria. And they would just be like, "Oh, it's labile." And they weren't connecting, they weren't going to the same conferences normally and so sometimes people would hear about it... the terminology really matters in connecting scientists, and I have no idea how people did things before the internet. Knowing about what other people were working on... it's crazy.
Forrest Meyen:All in-person conferences, I guess. Which is starting to feel like a thing of the past. Hopefully, we get back to those.
Jonathan:Virginia, I'm curious to learn more about... you had mentioned earlier in our conversation that maybe almost sort of coincidentally, or fate shined on you and maybe an advisor, like earlier in your academic career said, let's look at some of this particular field to study and that path led you to here. So, walk us back to what maybe... so like baby Virginia was, where are you, and what did you think you might want to do with your life?
Virginia Burger:Yeah. So I mean, there were a couple of key moments in getting me to intrinsically disordered proteins. One was, I was studying math. And I'd actually started studying biology. But all of the freshman-sophomore courses for biology were about like plants and animals. And I really just cared about microbiology. And I was like I don't want to doing this. And so then I switched to math, because the classes I really liked from my freshman year were the math classes. And so I was like, I'll just switch to math, and then eventually, I'll be able to get back to biology. And then I really didn't know what I was doing then, because I was studying math. And I kind of forgot that I liked biology. And so I started taking courses in computer science, just because I felt like computer games were really popular at the time. And that Sims was kind of the one game that went kind of for multiple genders generally, not that like.... Girls can like anything they want. But it was the world's most popular game, and it was the one that was not shooting
Jonathan:And it happened to be highly addictive so.
Virginia Burger:Yeah.
Malvika Miller:Gaming to protiens – I love it.
Virginia Burger:Yeah, so I was like, maybe with my math degree, I could actually be a game designer or something. Because I have this insight into games. Like, I didn't really play any games, but I knew that Sims was fun. And I was like, I bet as a female, I could invent other games that the other half of the population would enjoy. And so I started taking computer science courses. And then I really didn't want to do game design. So it was kind of like... but I was learning it. And then I went to a talk by a professor who's talking about the protein folding problem. And then I was like,"Oh, this is what I want to work on." And then he was like, "Oh, you've been doing all these programming courses. That's convenient." And so I started working—not with him, but with his collaborator. Because his collaborator, just had all this really great energy. And so we started working on the protein folding problem as my master's thesis. And I never went back to learning about game design at all after that. But it brought me this computer science skill that was very helpful. And then, when I was talking with him, like my college had a five year program, like you had to do a master's degree. So I was there, like fifth year, and he said, "What are you going to do next? Like we need to look for a PhD program or something for you?" And I was like, "Oh, I like this protein folding problem, like what else can I do with that beyond this?" And he very conveniently had dinner with his postdoc advisor the day before whose daughter had just gotten a faculty position teaching computational biology. And so he said, I think the field is called computational biology. And so then I went home and I Googled computational biology. And there were six PhD programs, and I applied for them. And that was how I started studying computational biology. And then when I was interviewing for computational biology, I went to Pittsburgh, and they had like a dinner where everybody went out during the interviews, which was really nice. And I was sitting next to this professor, and she told me about intrinsically disordered proteins. And I was like, "okay," and then I applied for... I didn't apply, but I accepted the offer from them. And I started working with her. And so it really just kind of like... if he hadn't known about computational biology, I don't know. I wasn't really sure what I would have done. So it just kind of worked out.
Jonathan:Well, I think Will Wright, one of the creators of The Sims, would be delighted to hear that story. An indirect path.
Virginia Burger:I think I played the Sims like a little that year of college and then once I started working on proteins, like it was just way more addictive.
Forrest Meyen:Proteins are way more fun.
Virginia Burger:That was fun, too. Yeah.
Malvika Miller:Maybe you can make a new protein game.
Virginia Burger:There are protein games.
Malvika Miller:There are? Wow, okay.
Virginia Burger:Yeah, there's games for solving protein folding that kids can play. Like there's a game called Foldit. It's really cute. Yeah.
Malvika Miller:Have you played that with undergrads?
Virginia Burger:No, when we had like mentees come in, we played it with them, because it's fun. Especially middle schoolers.
Malvika Miller:Seems like a good thing to put in your interview process. Finding some folks, how do you do on Foldit?
Virginia Burger:That would be a good question.
Jonathan:Something that over the years... I've done administering of distributed computing projects, and one of those was a Folding@home project and that over running over a year's time, sort of like Citizen Science where I have a computer that I can leave it on and connect to the internet. And it'll number crunch to help figure out protein folding and give it back to science. And so is it the case that this is still very much an open problem that science and scientists need help to be able to figure out the world of proteins: like we can't, right now, super accurately predict how every protein will behave? Is that the case or have we moved beyond that time, since my days of supporting the Folding at Home project?
Virginia Burger:I supported Folding@home too. So in the fall, Deep Mind, which is part of Google, their AlphaFold algorithm solved the protein folding problem for a lot of proteins. I think for proteins that crystallize, they can give an extremely accurate prediction of their structures, which is really impressive. And so knowing what they look like, and what the Folding@home was doing, I think is less relevant for proteins that are already in the crystal structure database, the PDB. But for proteins that are more flexible, there's still a lot of open questions. And then for proteins that can have multiple different folds, there's still a lot of open questions. And so it's definitely like, they've done a really an amazing thing, using these AI models to predict their structures. But there's still... I mean, 30% of proteins are intrinsically disordered. And so those regions, the way they're predicting it doesn't really help. And then knowing what the protein actually does is a whole nother problem that I don't actually work on. But I think it's... I mean, I'm sure people are making progress there. But it's a lot of different experiments to narrow down: like, what is this protein do? And I don't know if that's something we'll ever be able to just predict; at least not anytime soon, because there's so many different pathways in cells, and you have to narrow it down.
Forrest Meyen:So for you intrinsically disordered proteins, is the biggest challenge right now, like computational power, or is it getting that training data for the AI models?
Virginia Burger:I think it's the accuracy. And so right now, I would say where we are isn't the training data. We're good with our training data, but it's making sure for larger and larger systems that we're able to still model them accurately and then it will become a computing power problem because eventually we'll get to systems that are so large where simulations are just too slow. Because the simulation time kind of squares, at least with the number of atoms in the system.
Jonathan:So we have a sense of how you've gotten into the field and walking through sort of the academic corridors to that pathway. Where I would like to understand better is what got you as what I might characterize as sort of the academic Virginia Burger, reading, say, Elon Musk's blog on first principles and saying, and then using that as fuel to be like, let's start a company, because that's a big leap for so many folks to cross that kind of a chasm. So what gave you the courage and motivation to do this?
Virginia Burger:Yeah, so I was at MIT. And there's a lot of energy towards doing startups. And what really changed is... because I worked in startups during my PhD, like as an intern and done different projects for startups. But it never occurred to me to start a startup. I didn't even really know what they were. But then I was applying for faculty positions. And, I would go to seminars at MIT about what it's like to be a faculty member and kind of learn about it. And I was just hearing a lot about applying for grants and the challenge between finding time to work with students versus do research and teach and it sounded really nice and it was what I wanted to do, but I wasn't quite connecting with it anymore. And then one of my friends had a networking event she wanted to go to, and she needed somebody to buddy with her, and so I went with her at Sloan, the business school. And it was a networking event for people who were starting startups, and everybody was just so enthusiastic about what they were doing and was like "I need a partner for this" or "I need people to come work for my company." And I just really liked the excitement. And then I realized that like, "Oh, these are all like PhDs
and postdocs and business students too:there's no reason why I couldn't do this." And I think that had never been apparent to me until that seminar: that normal people can start companies. And so then I told my PI, "I want to start a startup." And he'd been so helpful. Like before faculty interviews, he would wake up at 6am and go through my slides. And he was just like, really supportive of me going forward and the faculty path. And then I was like, "Yeah, so we're not moving forward with me of that; I want to apply for startup grants." And so then he became supportive of that too really quickly. And so he had a friend who was a professor at Sloan teaching Healthcare Ventures. And so he said, "You should go take Healthcare Ventures." And then he was mentoring a master's student who was starting a company. And so he was like, "You should talk to her." And like, all of a sudden, there were all these other resources available. And MIT had so many programs, and I took them all. And they also had a biotech course. So I took their biotech course. And so my final semester was just all learning about startups and how to do a startup. And so that was what really made it possible: just signing up for these classes. And I didn't even like enroll. I just kind of walked in. And they were also welcoming, because the whole startup community is just like, "yeah, start a company." And so I was like, "Okay, I will." So it really was just like this one event and seeing this positive energy. And then l was really afraid I would like in a year lose the excitement, but I haven't. So I think it was the right decision.
Malvika Miller:That's really exciting. And yeah, my understanding from reading a little bit about the company is that you were incubated at Petri Bio. And I've learned a little bit about that model. And can you tell us like what that was like for you? How did you choose to sort of work with them? And when did you decide was the right time to really like pull the trigger?
Virginia Burger:Yeah. So they were announced in October of 2019. And I think I was kind of interviewing at a different accelerator. And then somebody told me about this one in Boston that was like, built around companies that are around bio and tech. And so I just was like,"Oh, well, why aren't I in Boston?" Because that's where all the biotechs are. And so I emailed and then I interviewed and then it took like.... I don't know, two or three weeks, and I was there. And so what their model is is they're not actually an accelerator, they're like a pre-seed program that provides a lot of support and mentorship for very early stage companies, which... kind of it's hard to find that because if you're not still inside your academic lab, and you don't want to build it with like university technology, but you want to build your own thing, there's kind of a gap there. And so I'd gone and worked for two years so that I would have like some funding to start the company. And also, I really wanted to learn what it was like to work at a startup and also how to do interactions with pharmaceutical companies and everything. So I did that for two years. But then you're still like, Well, how do I hire other people? How do I do all of this, and who do I even need and Petri really came in and made... just kind of like a step list of like this is how you start your company. And so it was really great, but they appeared right when I needed them. It's very convenient timing on their part. And I'd actually met Tony before through the... Tony's the person who is one of the founders of Petri, and I'd met him because he was teaching the biotech club or something at MIT. And so I kind of had known him through there, but I hadn't thought to follow up on the connection because he was interested in antibiotics at the time. And then it just kind of worked out. Petri has been really very, very great and supportive. And I think that model of letting anybody start a company is really important... of bringing in people with ideas and giving them the chance to move forward with them.
Malvika Miller:I think yeah, it's fascinating. You brought.. you'd had this academic experience that's super rich, and you've worked in top institutions and labs, and then you actually went to go work at a startup. Now you have your startup. What are the few like nuggets that you've pulled from a couple of those experiences that you think our viewers who were interested in early stage tech—what should they know? Like you've had these wide variety of experiences. So it's amazing.
Virginia Burger:Yeah. So I guess one thing that I've been thinking I would do differently is I would have an executive assistant from day one. The most important asset of the company right now is the founder's time. And I have put so much of my time into so many things that I'm
like learning on the fly right now:like, what is payroll? I don't know. Like, how do you do accounting? I have no idea. How would anybody know that until they've done it. And that takes up a lot of time that I could be spending reading papers, or talking to the team or just doing anything scientifically helpful to the company. And I think bringing on somebody who's an expert in actually getting the organization part of the company going is really important. And so that's something that I would... like, it's kind of hard because you have to pay an executive assistant, but I would find some way to do that. Like have somebody part-time supporting you from the start. That's one thing I would do differently if I started the company now. But I mean, things have worked out. But I think everything would just have been so much smoother if there had been somebody near me who said, like, "oh, you need to remember to pay taxes." Like, we always pay our taxes, by the way, but I'll remember on like April 14, and like, I've done it before, but like on, you know, in February, or whenever people are supposed to remember.
Forrest Meyen:And those Delaware franchise fees, those are always earlier.
Jonathan:Oh, yeah.
Virginia Burger:Luckily, the lawyers told me about those. So they'll email me reminders. But so like, I think, things that you could delegate, you should delegate as much as you're able to, is something that I've learned, but I think just all of the resources around Massachusetts have been so wonderful to actually have the company be possible. Like we took part in the MassCONNECT program, which gave me this huge network of mentors from pharma and legal teams, and just people with different areas of expertise around how you start these companies. And it was an amazing program. And we did the MassNextGen program for female CEOs, where they give us executive coaching, which is so helpful on how do you have difficult conversations with your team, how do you coach your team, how do companies even run? Like there's different types of management, and they've been kind of teaching all those things. So we've really benefited from all of the programs that are out there to benefit people: I take advantage of all of them.
Forrest Meyen:Awesome. Well, we're about at time. So we wanted to thank you very much for joining us on the show today. And we wanted to give you one last chance to, I don't know, give give your pitch to our audience of whatever else you want to pitch. I know that you are hiring them, but any other last thoughts that you have?
Virginia Burger:Well, we're definitely hiring people who are excited about drugging IDPs and who are excited about building new computational models and integrating that with experiments for being able to accelerate drug discovery. So people who really are passionate and then have a background in this area or could explain to me what else they could offer, we'd really love to hear from. And we really like reading cover letters, so people should write cover letters when they apply.
Jonathan: That's so nice:to like to read the cover letters. Virginia, I've heard that you have about three books that you've been working through: would you like to tell our audience about about what you've been reading?
Forrest Meyen:Tough Tech Today book club?
Jonathan:Yep.
Forrest Meyen:Everyone, this is your assignment between now and the next episode in two weeks.
Virginia Burger:I put them under my computer just now actually. One book that I absolutely love doesn't have a cover anymore. It's here. It's called Miracle Cure. And it's amazing. And it goes back on the history of antibiotics and just the whole course, and it's just so... I really like reading about the history of science. And I still really wonder how people did anything without the internet, as I said, and then this book is amazing. It's called The Truth in Small Doses. And it starts with this anecdote that people may know, but I didn't know about how, in like 1920 or so people thought that there were 24 pairs of chromosomes. And then, around 1950, a scientist finally stood up and said, "No, there's 23." And apparently for the 30 years before, people would kind of see it. And they'd be like, "Oh, maybe I miscounted. Maybe it's not always the same. Like, maybe... I don't know." People just took 30 years to say,"oh, that's wrong. There's actually 23." And I just like these books about like the history of science, and there's just so much of that where like, something happens and it's not what the technology can handle at the time. So they just ignore it. And I mean, like with IDPs, how like they were flexible regions of proteins, and they would just cut them off of crystals and say,"let's just not worry about this, yet." And I think, I don't know, I find that interesting and how to be more aware of that. And then I'm reading The Emperor of All Maladies, which is really good, but really sad. It's about cancer. So those are my books I brought for the book club.
Malvika Miller:Thank you for sharing.
Virginia Burger:Yes. Have you guys read those books?
Forrest Meyen:I have not yet. I have to put them on my list.
Virginia Burger:I highly recommend them.
Malvika Miller:No but I have The Gene from the same author of The Emperor of All Maladies. Yeah, but I haven't read the other two. Yeah.
Virginia Burger:Yeah, he's amazing.
Malvika Miller:Yeah.
Virginia Burger:Yes.
Jonathan:Well, Virginia, we have one final request. And that is could we always invite our guests at the end to say, like, "Hi, I'm Virginia Burger, CEO and Co-Founder of New Equilibrium." Ack, I can't say it but you can say it. "Stay tough." Could you do that?—name who you are and "stay tough?"
Virginia Burger:Yeah. Hi, I'm Virginia Burger, CEO and Founder and Co-founder of New Equilibrium. Stick tough.
