Drug Discovery with 1910 Genetics: Knowing Your Tools
The Harry Glorikian Show
- Drug discovery involves the identification of a target molecule or pathway involved in a disease process. The process utilizes various techniques such as computational modeling, high-throughput screening, medicinal chemistry, and preclinical testing.
- Extensive testing in both animal and human subjects is conducted to assess safety and efficacy before approval for clinical use.
- Technology, including machine learning and AI, is increasingly being integrated into the drug discovery to enhance efficiency and effectiveness.
- Funding opportunities and startup culture have evolved, providing more support for entrepreneurs in the biotech industry.
- Successful drug discovery companies often secure funding through accelerators and venture capital firms.
- Partnerships with large pharmaceutical companies can be formed to leverage resources and expertise in developing drugs.
- Drug discovery companies may focus on specific therapeutic areas or diseases, and they build pipelines of potential drug candidates for internal development and external partnerships.
- The drug discovery process involves hit finding, hit-to-lead, and lead optimization stages for both small and large molecules.
- The ITO platform starts with identifying the problem and the available input data.
- Biological and computational data streams are used, with computational data often compensating for the low data regime in biotech.
- The data goes through multiple transformations using various model architectures.
- The transformed data is then subjected to further computational structural biology and synthesis.
- The output stage involves testing the synthesized molecules in in-house assays.
- The importance of quality control checkpoints is highlighted to ensure data integrity throughout the process.
- Jen Nwankwo emphasizes that the choice of model architecture should be based on the scientific output and relevance to the problem rather than relying on specific models as the most important decision.
- The project exemplifies the application of the Input, Transform, Output approach in defining the problem, identifying the necessary data streams, and leveraging multiple model architectures.
- They aim to highlight the potential of their platform to tackle both small and large-molecule therapeutic questions and promote cross-learning opportunities and announce partnerships and share more about their work in the coming year.
Harry Glorikian: Hello. I’m Harry Glorikian, and this is The Harry Glorikian Show, where we explore how technology is changing everything we know about healthcare.
You know me as a podcast host. But in my other life as a life science investor, I often run into people who’ve been on such interesting, compelling personal journeys, or who have such an inspiring way of looking at the world, that I know I want to interview them for the show, almost no matter what they’re working on.
I knew Jen was one of those people when I first ran into her two or three years ago. But Jen isn’t just a compelling personality. She’s also the founder and CEO of a fascinating drug discovery company here in Boston called 1910 Genetics.
Jen’s PhD is in pharmacology. And that shows through in her practical focus on fixing the drug discovery process to get more and better therapies into the hands of doctors.
To hear Jen tell it, 1910 Genetics is focused on finding the most promising new drug discovery candidates for stubborn health problems, and it takes a refreshingly agnostic approach to everything else.
The company doesn’t hunt for just small-molecule drugs or just protein therapies. It explores both. It doesn’t utilize just one form of neural networking or machine learning. It uses whatever model produces the best science for a given problem. It doesn’t hunt for drugs using just wet lab data or just computational simulations. It does both. It isn’t just assembling its own pipeline of drugs or just partnering with larger pharma companies. It’s working on both.
Jen wasn’t even dead set on being an entrepreneur—she had to be talked into applying to the Y Combinator startup incubator and into accepting her Series A investment from Microsoft’s venture fun.
She says the way 1910 thinks about drug discovery is to start with the desired output — say, a new molecule to block pain — then figure out what sorts of data inputs exist. Then they find or create all the data they need to analyze the problem. Then they transform that data using whatever AI tools work best, until they get some decent drug discovery candidates.
She calls it Input, Transform, Output. It’s never that simple, of course. But at a time when AI and machine learning focused drug discovery companies are sprouting up faster than dandelions — each one touting some specific reason why its model is better than all the others — 1910 Genetics is has a more inclusive approach to solving classic problems in pharmacology, and it’s one that I’d love to see spread to other parts of the life science business.
In our interview back in February, we got a chance to talk about Jen’s path from scientific success to starting a company, 1910’s business model and drug discovery pipeline, where the name 1910 itself came from, and a bunch more.
Here’s our full interview.
Harry Glorikian: Jen, welcome to the show.
Jen Nwankwo: Thanks for having me, Harry. I’ve been looking forward to this one.
Harry Glorikian: This is a long time coming. I mean, we’ve been talking about you being on the show for, I don’t know, at least two years, but it’s good to finally have you here. And I’m sure that, you know, two years later, there’s so much more to talk about. But let me step back here for a second. Right, because. When you look at the drug discovery process today. You know what’s wrong with it? What aspects of it have you set out to fix at 1910 Genetics? Just sort of from a high level.
Jen Nwankwo: Sure. The drug discovery process today, I’d say there’s probably maybe three main things wrong with it. The first is that it takes too long. On average, you’re looking at about 12 years to bring a new drug to market for most diseases. Could be shorter for rare diseases, but on average, 12, 12 years. It costs a lot of money, upwards of $2 billion each time for each therapeutic. And, you know, it has a very low probability of success. You know, only about 10% of drug discovery candidates that begin a first in human studies would end up making it all the way to FDA approval. And if you go all the way back to the discovery stage, the odds are even lower. You’re looking at like some 0.00% or something like that. So, you know, timely, very lengthy, very costly and a very risky endeavor. Those are the things that, say, are wrong with the drug discovery process today. And the part that, you know, I founded in 1910 Genetics to tackle is really the entire thing, but really started out with the discovery phase first and kind of like building tools and technologies, you know, along the value chain. So starting out focusing on that initial pre-clinical drug discovery phase, you know, that encompasses target selection, hit finding, hit to lead optimization, finishing out some preclinical work and then before first in human studies. So our goal is in fact, to impact all aspects of the value chain. But we have started on the discovery side.
Harry Glorikian: Right. So I guess at a high level, what are the core technologies you’re working on at the company? I mean, I’d like to come back and dive into like more details in a few minutes, but if you can sort of give the listeners a general picture of the, let’s say, the unique capabilities you have at 1910 Genetics, right?
Jen Nwankwo: You know, today there’s a lot of talk about generative AI with ChatGPT. So I’m sure even a layman has heard about artificial intelligence and so on. And there’s certainly lots of companies, even on this show of yours, who have who you’ve talked to and are taking different approaches to leveraging AI and drug discovery. What I will say that differentiates 1910 Genetics is that we are, as far as we know, the only company that has built a synergistic, dual purpose, small and large molecule platform that integrates artificial intelligence with computation and with wet lab biological automation to design drug-like molecules better, faster and cheaper than traditional approaches. So in terms of, you know, what are the core pieces of the technology, for us, it’s that sort of three legged stool of, of the artificial intelligence, the computation and the wet lab biological automation, bringing three of them together and deploying them in a synergistic manner and in a modality-agnostic manner. I think the modality agnostic manner is perhaps our biggest differentiation. Where all the other companies in the space are either tackling primarily small molecules or they’re doing primarily large molecules. And what we’re saying is we see synergies between these two modalities that we have trained our technology to not treat them as, you know, completely different, but to actually cross learn from them and to take learnings from small molecule discovery and use that to be better at large molecule discovery and vice versa.
Harry Glorikian: Interesting. So. I guess maybe now’s a good time to step back for a second and say 1910 Genetics, what is it? Tell us the story behind that. I mean, I think sickle cell had you know, is is one of the underlying reasons for that 1910 Genetics name. But maybe you can give people an idea of how you came up with that name.
Jen Nwankwo: Yeah, You know, it was, it was just, I was just trying to have fun. First of all, I was trying to have fun with the name. We’re not like, you know, a dedicated sickle cell disease company. That’s the first myth I want to dispel about us. And the name comes from 1910 is the year that the first patient in the United States with sickle cell anemia was diagnosed. That is true, but what relevance does sickle cell disease have for us? Well, sickle cell is the first disease for which we completely understand the molecular basis, where you can look at a disease and you can say, I know exactly down to the molecular level what causes this disease. And we think of that as the North Star for the types of diseases we want to go after. We want to enter therapeutic areas, indications where the biology is really clear. We to the to the extent possible, don’t want to take biological risk. Right. You know, like if you look at a space like Alzheimer’s, for example, you know, does beta amyloid cause Alzheimer’s? Is it a cause? Is it an effect by biology? Like, that is uncertain, is not exciting to us. And so when we started out the company, we said we want to focus on the molecule we want to take risk around, can we design a molecule we don’t want to take risk around? Do I understand the underlying biology driving this disease? And so we said we’re going to use sickle cell as an example of the types of diseases that we want to go after, where there’s a clear causative driver of the underlying biology. And so 1910 is meant as a North Star for us because we realized that, you know, the best molecule will not help you if you’ve got the wrong biology. And so that’s that’s sort of like how I came to the name.
Harry Glorikian: So, you know, a lot of times we talk about people’s career journeys, right? And so, you know, I want to just dig into that just a little bit. Right? But and I want to share a little bit of your background with the audience. And so if I get something wrong, you you can sort of correct me, but I know you studied biochemistry in college and then went on to your doctoral research at Harvard Medical School and Tufts School of Medicine on blood, platelets, sickle cell disease. Coincidence. We were just talking about it. And the chemical pathways behind pain. Right. And then I believe you were a Howard Hughes Medical Institute fellow, both at Tufts and at Boston Children’s Hospital, where you focused on sickle cell drug discovery. And then you made the leap into the business world where a few years later you were directing business development for a life science startup called Transparency, which I know a little bit about. And then you did management consulting in the health care industry for Bain and Company for a little while. And then finally in 2018, if I correct you, founded in 1910 Genetics. So if there was a single passion that guided you through all of those experiences, how would you describe it to people who are listening and and in the end, what sort of mean drove you to the field of computational drug discovery.
Jen Nwankwo: I’d say, first of all, fantastic summary. Thank you. You got the most part. I did do my undergrad in biochemistry. My PhD was in pharmacology broadly as a classical sort of discipline. And I set in, I set out to study just what I call classical pharmacology, which is basically the art of, or the science and art of, of designing drugs and like, what does it take to make a therapeutic. And you know, I was, I applied that training in a variety of different places. Hematology was one of them. So platelets, platelet physiology, platelet applied to sickle cell and other sort of thrombotic conditions. That was one. But there was a lot about my PhD work that was not published. I did a lot of work, for example, in diabetes and hypoglycemia, and I did a lot of work in some other areas as well. So it was a PhD in pharmacology from the Tufts University School of Medicine. It was funded by the Howard Hughes Medical Institute. I was fortunate to be selected for a competitive fellowship by HHMI, so I was a Predoctoral Fellow throughout my work. The work was done in part at the Boston Children’s Hospital and Harvard Medical School. So Harvard does get some credit as well for the innovations and the advancements made. As part of my my research and what I would say motivated me to found 1910, the first thing I’d say is I was not one of those kids that grew up like with a lemonade stand, you know, and had like aspirations to like, be an entrepreneur. Like, no, I don’t have one of those stories for you because growing up, as I look at a lot of the essays that I wrote as a kid, I never once said I wanted to be an entrepreneur.
Jen Nwankwo: I mean, I wanted to first of all, I wanted to be Jennifer Lopez. When I look at my essay, that was certainly an aspiration and then for a while, I wanted to be a pilot. And then for a while I wanted to be like the president of a country, you know? So I look at all the things I wrote throughout, like elementary school and even middle school, and there wasn’t a single “I want to be an entrepreneur” there. So I’m definitely one of the entrepreneurs, I’m definitely in the category of, like, entrepreneurs are made, not born. I don’t think I was born an entrepreneur. I think I sort of a combination of factors and obviously faith led me there. If I were to say what motivated me and what made me take the leap, I think founding 1910, a couple of things had to come together. The first one was I, as I mentioned, I trained as a classical pharmacologist and I learned how to use traditional tools to identify disease targets, to try to drug them and try to dose targets and all of that. And along some somewhere along the line, I sort of started to learn about, you know, predictive analytics before it was called AI/ML. Just the idea that you can learn from like, you know, data to make better informed decisions and that you could perhaps reduce the number of animal testing that you had to conduct in order to get some sort of a verdict on a drug discovery candidate that you were working on.
What is drug discovery?
Drug discovery is the process of identifying and developing new medications to treat or cure diseases. It involves the identification of a target molecule or pathway that is involved in a disease process and the development of a molecule or compound that can interact with that target to modify its activity. The process typically involves a combination of computational modeling, high-throughput screening, medicinal chemistry, and preclinical testing to identify and optimize potential drug candidates. Once a candidate has been identified, it undergoes extensive testing in both animal and human subjects to assess its safety and efficacy before it can be approved for clinical use. Drug discovery is a complex and expensive process that can take many years and involves collaboration between scientists, clinicians, and regulatory agencies.
Jen Nwankwo: And so I was sort of like keeping an eye on, you know, trends and how the field of machine learning was progressing. And I saw that there was perhaps an opportunity to bring it to solve a pain point that I had as a classical pharmacologist. So that was where that was what I’d say piqued my interest. But it wasn’t sufficient to go found a company because founding a company is a mental and emotional decision more than it is anything else. And I’ll say the second thing that sort of had to come together was I had to experience startup culture myself to be certain that I wanted to go do it. And I know you alluded to transparency, life sciences, and I had the good fortune to work with Thomas and Mark and to see them as two very experienced entrepreneurs starting a company at a time in their careers where honestly, they were really set. Right. They had, you know, they weren’t like young first time founders. They had, you know, been working in their respective industries for for decades. And to see them sort of like try to go from like, I’m going to put this together, we just my own like angel money to I’m going to try to raise institutional capital to I’m going to try to commercialize this technology to different pivots in business models. I enjoy that ride, and I had to experience it for myself for the greater part of three and a half years to sort of solidify in my mind that this was something that I wanted to to do.
Jen Nwankwo: I think while also transparency, you know, seeing that grand vision that the founders had to bring technology to the later stage of drug discovery and development, particularly clinical trial design, clinical trial execution and so on, I thought, you know, technology is going to have its moment in pharma. And I thought perhaps I could start a company to bring technology to the earliest stage. Right? I’m more of a bench scientist. I’m more of like a discovery person, not so much a development person. So I think that those things had to come together. And I think another thing that had to come together was a credible path to funding for first time entrepreneur, you know, and at the time, you know, accelerator programs like Y Combinator and now several others spinning up like Petri. Even just this past week. So many different, you know, vessels for funding for first time young founders, which hadn’t existed in biotech. You know, biotech had funded a certain phenotype of individual for for the better part of a century. Right. And so all of these things had to come together. And frankly, I had to sit and talk to my husband and say, look, I’m going to leave my very lucrative job and I’m going to make this leap. So so so there were some personal factors there. There were some macro factors there. And just had to, they all had to come together at the right time to actually go start 1910 genetics in late 2018.
Harry Glorikian: Yeah, well, you know, you mentioned Y Combinator like. You. You were part of that accelerator in the winter of 2019, like. What was the pitch? Right. Because whatever it was, Sam Altman liked it. And you know, made you part of the team. So what did you what did you say back then that got everybody’s attention?
Jen Nwankwo: Yeah, My I had a little bit of a non-traditional path to YC. The first the most important thing was I didn’t want to apply to YC. I did not want to do YC. I heard about it. I happened to be in San Francisco and they were hosting a female founders summit in late 2018, September 2018. And so I just went out there, frankly, for free food and free wine. And, you know, I met all these people. They sounded smart. And then I told them about what I was, my background. And just as a founder and like what I was thinking about. And I just went back to Boston and in the deadline came for the YC application. And I did not apply. I really was not going to do it. And then I got an email from like three different partners at YC emailed me to say, Hey, we didn’t see an application from you. We thought that you were very interested in the program. You know, what can we do to encourage you to apply and so on and so forth. But long story short, they finally encouraged me and convinced me to to apply. And so I did. And I got in and I participated in the winter 2019 batch, which went from January through March, April of 2019.
Jen Nwankwo: Speaking of speaking of Sam Altman, Sam did not initially invest in us going through the batch. He invested after the batch as part of our as our seed round. He ended up making the largest single investment into the seed round. So he is the lead of the seed round. And Sam came in, uh, precisely during COVID when, um, you know, there was a lot of interest to apply, you know, just accelerated approaches to try to come up with different therapeutics for, for, for COVID and for us, we were able to deploy our platform against a couple of host proteases. So, human host proteases that help the SARS-CoV-2 virus enter the cells. And we were able to very rapidly identify some promising initial lead molecule candidates on the small molecule side. And that work was featured in Science magazine and and the NIH and NIAID and NCAS invited us to to talk about that work and how we’re able to deploy so quickly to come up with some leads. We ultimately, just from a strategic perspective, did not choose to continue further development of some of these starting hits. But that was where Sam came into into the picture to, to to sort of round out the syndicate in our seed round. And I think the pitch for us at the time was that we were going to, you know, leverage sort of different biological pathways for which I had intimate knowledge, for example, like the Calpain pathway, cysteine protease signaling, calcium signaling pathways, dysregulation in calcium homeostasis, more broadly, and how it impacted different kinds of diseases across the neuroscience, oncology, immunology spectrum. And we’re going to find different ways to drug them. I think that we’re very much still doing that today, but the vision for 1910 today is far bigger than dysregulation in calcium homeostasis and all the diseases that result from it. So I’d say if you looked at our application, it was, it was, it was what we are today is probably 100 X of that same idea, but we have not deviated too much from this underlying concept of we want to understand disease biology. We want to understand disease biology, and then as we are growing as a company, we’re building more and more innovative tools to sort of tackle disease biology.
Harry Glorikian: Yeah. And so, you know, just to give everybody sort of the stages. Right. You were at YC, you came back, you set up at Lab Central because I remember I came and visited you there. Then, um, well, you identified these two molecules.
Jen Nwankwo: We actually didn’t come from, so when we came from YC, we set up first at Cambridge Innovation Center. We did not go into Lab Central directly. So for about a year or so.
Harry Glorikian: That’s right. We talked about. We sat down and we talked about that.
Jen Nwankwo: Cambridge Innovation Center. And while we were at Cambridge Innovation Center, we were evaluating different incubators. So applying to places like Lab Central, Alexandria, Launch Labs and so on. And we were accepted at Lab Central and then moved into Lab Central in November of 2019. Yeah.
Harry Glorikian: Yeah, I remember like we got together and it was funny. And this is a strange story, and maybe it’s dating me, but I remember you said, you know, you’re sort of woke. And I was like, I don’t know what that word means. I should go look that up. But so. You. You found these two molecules in coronavirus. I think you made a you know, you pitched it at a virtual summit that Dr. Anthony Fauci had convened. And I think, if I’m not mistaken, that’s when you made your first contact with Microsoft, which eventually became the lead investor for your series A. And so just to tell everybody like, I’m fast forwarding to 2021, you did a $22 million Series A round. You moved your offices to the seaport. Let’s talk about where the company is today. I don’t know how many people do you have? What are the main programs you’re working on? What’s the business model? Are you collaborating with large pharma? I mean, give me the the you know, let’s get people up to date on where you are with the company and what’s going on.
Jen Nwankwo: Yeah, just a little bit of a summary. Again, started the company in May of 2018, did a $4 million seed round that was led by Sam Altman. That round closed in 2019, early 2020, and then middle of 2021 or early 2021 rather did another $22 million in Series A that was co-led by M12, Microsoft’s venture fund, and Playground Global. And in terms of how Microsoft came to learn about us, just a quick story. In late 2020, one of my friends that I had known from when I was a PhD student at Tufts, Koki Hirosaki, he was a research investigator at Novartis when I was a PhD student at Tufts, and we invited him to give a seminar. So our PhD student group, and that was where Koki and I met each other for the first time. And we wouldn’t have, we couldn’t have imagined that ten years later, Koki would be an investor in 1910 because he went on to join Microsoft’s venture fund and he was actually the one who sourced 1910 and actually convinced us to begin raising a Series A when we actually didn’t need the money and weren’t we were not in a fund raising process. Microsoft preempted the round. But all that said, we closed that additional $22 million financing in February of 2021. And the goal was to, you know, build out the team.
Jen Nwankwo: At the time of the series close we were like seven employees, myself and like six other very brilliant but young and fresh graduates from PhD programs, top PhD programs, you know, with backgrounds in like computation and some biology and AI and so on. So a key goal of the Series A was in fact to build a management team to, to build, to bring on people who actually knew how to develop drugs. I’d say a second goal of the series A financing, was to take the AI and chemistry platforms that we had built and scale them up. At the time of the series A funding we were doing, like, we were able to design drugs for like one target at a time, and we wanted to scale up the and parallelize the platform such that we could be working on 10 targets in parallel if we needed to. So there was a lot of scale up that needed to be done to the platform. The third thing that we had to do with the series A funding, was to build out our own biological wet lab capabilities. It’s a it’s a key differentiation for us as well. And there’s a lot of companies that are computation only or virtual companies and so on. We recognize early on that our ability to control our fate on the biology side, at least on the early biology, was going to be key.
Jen Nwankwo: And so we sort of spent like 13 months building out our own facility in the seaport, outfitting it and then and automating it and scaling up our assay design capabilities more than 1000x such that we can continue to feed the AI and ML with high quality proprietary data sets. And I’d say the fourth goal for the series A financing, was to combine all of those scaled elements. So skilled team, a scaled AI platform, a scaled wet lab platform to bring all of that together and start translating it into a drug discovery pipeline. We were very, we were very intentional in the beginning in recognizing that there was going to be a period of time where we were just focused on the platform build and we didn’t want to be too distracted by, you know, sort of like traditional asset discovery because we knew that if we got the platform elements right when we begin deploying it, it would be easier for us to move assets forward. So we we got to a point last year where we started to think about, okay, I think the platform is has scaled considerably. It’s not all the way where it needs to be, but has scaled considerably for us to start, you know, growing the drug discovery pipeline.
Jen Nwankwo: And at that point, we we sort of set out a business model, as you alluded to. On the one hand, we wanted to build our own internal pipeline of a few programs that we want to take all the way to FDA approval, hopefully ourselves, with subsequent rounds of fundraising and other sorts sources of non-dilutive funding. And then on the other hand, we wanted to create an external pipeline that was that would be comprised of programs that are partnered with pharma companies, et cetera, whom approach us to leverage our platforms in discovering drugs against targets of their own interest. So last year we made strides on both fronts. On our internal pipeline, we clearly define our therapeutic area strategy in areas like neuroscience, which includes neurology, pain, and neurodegeneration. But this sort of new burgeoning area of precision neuroscience, which means you want to do for neuroscience what was done for oncology some 20 years ago. And then the other area we defined was immunology, autoimmune. And then rare and genetic diseases. We prioritize these three and we have we had programs in all of these three areas moving at different stages in the early drug discovery phase with the most advanced being in lead optimization today. And then on the external partnership side of things, we did our first sort of initial partnership with a major pharma company in the autoimmune immune space and kind of use that as, you know, a test case to work out a lot of kinks and everything from like IP arrangements to economics, how we structure deals, because, you know, partnering with pharma obviously is very important.
Jen Nwankwo: It’s a very critical, you know, source of economics for a lot of biotech companies. But for platform companies in particular, thinking about the business model now versus later, thinking about the financial structure of deals is very important. So we were really able to use that first partnership to work out, you know, how should we do deals going forward, you know, and that was beneficial. And so this year is, is what I call our partnering year. We brought on a chief business officer at the end of last year because partnering is a key focus for us this year. We feel like we have scaled the platform enough and we have more clarity around the business model and around what types of partnerships we really want to be in and honestly what kinds of partnerships we don’t want to be in. And so this year we’re only two months into the year, but we have several ongoing partnering conversations and we’re hoping to close on a number of those in the in the second quarter of the year.
Harry Glorikian: Yeah, I mean, I’ve talked to I was just talking to Alex from Insilico just earlier this week and he was saying like, we need partners. Like we can’t there’s no way for us to take everything that we’re the platform is generating forward. So, you know, Big Pharma is sort of our friend in a sense as opposed to a competitor, right? So when you’ve got a platform. You. You can’t take everything forward, right? You need somebody else to take some, you know, promising products forward for you or with you.
Harry Glorikian: Let’s pause the conversation for a minute to talk about one small but important thing you can do, to help keep the podcast going. And that’s leave a rating and a review for the show on Apple Podcasts.
All you have to do is open the Apple Podcasts app on your smartphone, search for The Harry Glorikian Show, and scroll down to the Ratings & Reviews section. Tap the stars to rate the show, and then tap the link that says Write a Review to leave your comments.
It’ll only take a minute, but you’ll be doing a lot to help other listeners discover the show.
And one more thing. If you like the interviews we do here on the show I know you’ll like my new book, The Future You: How Artificial Intelligence Can Help You Get Healthier, Stress Less, and Live Longer.
It’s a friendly and accessible tour of all the ways today’s information technologies are helping us diagnose diseases faster, treat them more precisely, and create personalized diet and exercise programs to prevent them in the first place.
The book is now available in print and ebook formats. Just go to Amazon or Barnes & Noble and search for The Future You by Harry Glorikian.
And now, back to the show.
Harry Glorikian: So let’s jump back to the technology at the sort of the core of what you’re doing. You’ve got a single discovery platform which can accelerate both small and large molecule drugs. Maybe we start with the small molecule platform. Like can you explain, you know, what does the discovery platform do, what makes it unique. Maybe you can break down the process a little bit from hit discovery to lead generation to lead optimization. You know, I don’t know where does the training data come from for the machine learning algorithms? You get the idea sort of. Right. Draw a picture for us if you can.
Jen Nwankwo: Right. The first thing I would say is that we we don’t see the clear distinction between large and small molecules that a lot of people see. And I know that probably sounds like blasphemy. So most of the industry … so what I’m just going to say to you is that the processes or the stages in early drug discovery for both small and large molecules are the same. It’s you start out with hit finding, you want to find a hit compound or a hit large molecule. Then you go to hit to lead. Then you go to lead optimization. At least those three stages, both large and small molecules. Go through those same stages. And what we have built is a single platform that works and we call it Input, Transform, Output or ITO. And so given a problem—and we always start with the problem, which is the output—given a problem, a drug discovery design problem, we begin by asking ourselves, what is the input data we have? So a drug discovery design problem could be, Hey, here’s a kinase. I want you to design a novel allosteric selective inhibitor of this kinase using this as a small molecule. Okay. Or you could say I am trying to, I have a current antibody and I want to increase its half-life from, say, four weeks to 12 weeks. At the surface those are two very different questions pertaining to two very different modalities. One is around creating a new molecule from scratch. The other is around optimizing existing antibody. And what we’re saying is at 1910, we see similarities in both of those questions.
Jen Nwankwo: And we will go through the same process of asking, let’s start with the input. What is the, what are the sets of input data that we have, existing data. And then we dive into the biological data stream. We talked about having our own high powered biological automation facility. So what data can we generate at scale? And then the second thing we go to is computational data stream. Big Pharma is characterized by a low data regime. All of biotech is you never have enough data to leverage certain types of technologies. And one of the ways that we are overcoming the low data regime problem is by supplementing biological data streams with computational data streams. But think about computational data streams is that while you can generate them at scale, they tend to have a lot of noise. And so you really need to figure out like computationally, what computational data is actually relevant to the disease, to the clinical endpoint and so on. So we take all of this and then we begin the process of transforming it. We do multiple levels of transformation and this is where the AI is doing the transformation, transforming this data, computational only, biological only, the mixture of computational and biological and so on and so forth, continuing to transform it with multiple different model architectures. So where generative AI is the right one to use, you use that where it’s not the appropriate one to use, you pick the right architecture for the problem and you’re not, you’re not trying to just throw the latest and greatest at a problem when it doesn’t fit.
Jen Nwankwo: And so that’s what we would say, transform the data. What I mean is, is the process of model selection, architecture, selection, data curation data, data prep, data, featurization, data representation and getting to see which models are doing the best job at answering that question. Remember, the questions were very diverse, and so once we’re done with the transformation process, we do further computational structural biology, and then we’re going through the process of synthesis. So in the case of antibodies, we can synthesize the resulting antibodies and see whether they have increased half life or we can synthesize a small molecules and see whether they’re in fact novel and selective and so on. And then we get to the output stage and we begin testing them in our own in-house assays. We have developed primary and secondary biochemical and cellular assays at scale where we can test. And so we come out with the output. And so what we are essentially saying is that there are a lot of problems in pharma that in drug discovery that you can use the same Input, Transform, Output single process to tackle. And all you just need to be clear about is what are your data sources? How do you overcome the low data regime problem? How do you transform the data and are you transforming the data in the right way? And how do you think about the correct output metric that actually is relevant to the disease to to the indication? Et cetera.
Harry Glorikian: Well and all the, you know, the quality control checkpoints along the way. Right. To make sure that your data is as you’re transforming you’re not sort of, there’s no errors being introduced in that process. Right? So from an engineering perspective, is just setting up ways to say, yes, this is running the right way and the data coming out looks. Correct?
Jen Nwankwo: Right. Right. Well, there’s the machine learning engineering piece, which is around things like machine learning, operations, model metrics, performance runtime and all of that. There are those things. But what I mean when I say model selection is just the performance, the scientific results that you’re getting. If, for example, you had tasked the generative AI with, um, in the transformation step, you had given it some input data and you wanted it to transform that input data and generate new molecules for you, you are not only going to be looking at the model metrics of like model runtime and so on, you’re going to look at the actual molecules that were generated and you’re going to ask, well, are these molecules even valid? Are these drug-like? Are these even molecules I want to bother manufacturing. So when I talk about looking at the output of the models, I’m actually talking about the actual scientific output that you’re getting and asking yourself, is this the right model architecture based on the output I’m getting? Is this the right model architecture for this particular problem, or do I need a different tool out of my toolbox? And the good thing about the way we approach it is we don’t try to sell you on any particular architecture as opposed to perhaps some other companies who would say, Oh, I’ve built a graph convolutional neural network that does this particular thing. We for us, it’s not about the model itself because the models, frankly, are commoditized, right? They’re commoditized. Everybody has access to graph convolutional neural networks or, you know, some form of reinforcement learning or whatever. So I often take a step back when I hear companies begin their pitch with like, I have a graph convolutional neural network that does this because I’m like, There are so many times we have run graph convolutional neural networks and the scientific outputs were not what we desired, you know what I mean? And so why, why sort of hone in on the model type? Because that’s rarely ever the most important decision you have to make on the way to discovering a new therapeutic.
Harry Glorikian: Yeah. And you can imagine from where I sit, hearing all these pitches, at some point you’re like, Oh, Jesus. So but, you know, I did find maybe we could talk about one of your projects because I did find some stuff that sort of already, you know, putting publicly out there. I think you got a $550,000 grant from NIH to look at small molecule non-opioid drugs for the treatment of chronic pain, which would be awesome. But, you know. How does that problem lend itself to the 1910 approach? I mean, what do you know beyond that? What have you learned from that project about maybe the genetic roots of pain or, you know, or about the best ways to use computational methods to sort of get to what you’re trying to, to uncover in this project.
Jen Nwankwo: Yeah, I know there’s a lot that I love about our ongoing work that’s funded by the National Institutes for Neurological Disorders and Stroke, the NINDS and the NIH. First and foremost is it builds on disease biology that I illuminated during my PhD work. So this idea around calcium dysregulation and how certain, um, you know, disease targets, how certain biological targets that are responsive to calcium, um, how they could, you know, you know, sort of start certain disease processes. So the underlying biology was something that I had uncovered during my PhD work. Now what I did not have during my PhD work was a novel molecule to, to sort of target that disease biology. And that’s where the NIH funding came in. And so the goal there was to can you again, going back to the 1910, Input, Transform, Output, single platform process is defining the output as, hey, we want to arrive at a small molecule, brain penetrant, selective protease family selective inhibitor that was also selective against the entire opioid family and could have broad application in different types of chronic pain. So whether it be neuropathic pain or, you know, um, you know, some forms of chronic pain associated with sickle cell disease or, you know, peripheral inflammatory pain that is chronic in etiology and so on. And so we got that grant from the NIH about 18 months ago, and, and we’ve been working on that project and we actually now in lead optimization with some of our lead series.
Jen Nwankwo: So again, it’s about we don’t think of that problem any different than, say, the antibody problem I described to you. We again, to us it’s about what is the output, define the output, what are the metrics, and then work backwards from the output and say, okay, well what does my input need to be? What are the existing data that I have? What are the biological data streams I need to create the computational data streams I need to create? Then go to the transform step. How can I transform this data and what are the right sorts of approaches I need to use? And you might not know the right one initially, but you try a couple of different transformational steps. And in the case of this opioid example, we leveraged four different architectures. We leveraged a long short-term memory network, we leveraged a random matrix discriminator, we leveraged a graph neural network, all in parallel to to get to the final sets of initial lead molecules that we had today. So it’s an example of how we’ve taken a very clearly defined problem, turned it into a very clearly defined output in the form of like desired chemical characteristics in a lead molecule and a lead compound, and then started to build the input data that we need and transforming that data in all kinds of ways to arrive at what we hope to be hopefully a development candidate uh, soon.
Harry Glorikian: So, trying to get to the to the closing here. But as the company grows, what milestones should people will be watching for over the next, I don’t know, year, five years from the company? You know, how will you know or how will people be able to sort of recognize them? I know. I know. I’m always encouraging you like press release. We need, you know, information. Get information out there. Right? Yeah. Yeah. And I know. I know you like to keep, you know, stuff close to the vest. So how do we measure? How do you measure success?
Jen Nwankwo: I think for me, there’s the bigger question of how do I measure success of the company? And for me, I’ve always said I started this company to bring drugs to patients. And when all is said and done, that’s how I’m going to measure success. How how did 1910 succeed and in how many different indications did we succeed? So that’s sort of like my bigger picture, what I consider my life’s work, right. In terms of this year, what what people can look to for us is, yeah, we’ve, we’re not we’re not too big on the press releases as you’ve known that about me from the from the beginning. But I think this year we want to be a bit more vocal about technological progress. We’ve shied away up until now from publishing key aspects of our technology. But I think this year you’re going to see more from us, especially around this idea and this differentiation pillar of building a synergistic platform that can tackle both small and large molecule therapeutic questions. I think whether in the form of scientific publications or in the form of blog posts or scientific presentations, I think that this year we’re going to be more vocal than we’ve been in showing why we think the industry needs to start to think of these as being opportunities to cross learn, versus I’m a small-molecule guy or I’m an antibody guy, you know, and so on. I think that’s one of the big things that I think I’m excited about to see my team sort of be out there in different forums to sort of share how we’re thinking about this and why we think it’s it’s it’s somewhat revolutionary in the way drug discovery is being done. I’d say that potentially want to also be vocal about some partnerships this year we’ve we’ve done like I said, a major partnership in the past and we didn’t announce that. But I think this year we want to be vocal about partnerships. And this year we want to also be vocal about, you know, progress in our pipeline and ultimately be vocal about fundraising, whatever the next fundraising milestones for us might be.
Harry Glorikian: Well, as someone in the space, I look forward to reading the papers and reading the press releases and and, you know, keeping track of the company. And I wish you, you know, incredible success because, you know, we need companies in the space to move forward and be successful. So great having you on the show.
Jen Nwankwo: Thank you so much, Harry. It was my pleasure.
Harry Glorikian: That’s it for this week’s episode.
You can find a full transcript of this episode as well as the full archive of episodes of The Harry Glorikian Show and MoneyBall Medicine at our website.
Just go to glorikian.com and click on the tab Podcasts.
I’d like to thank our listeners for boosting The Harry Glorikian Show into the top two and a half percent of global podcasts.
To make sure you’ll never miss an episode, just open Apple Podcasts or your favorite podcast player and hit follow or subscribe.
And don’t forget to leave us a rating and review on Apple Podcasts.
We always love to hear from listeners on Twitter, where you can find me at hglorikian.
Thanks for listening, stay healthy, and be sure to tune in two weeks from now for our next interview.
FAQs about Drug Discovery
I like to include small FAQ sections at the end of my podcast transcripts to give the readers a better idea of the topic. So, here are the FAQs about drug discovery.
What is the drug discovery process today?
The drug discovery process today typically involves the following steps:
- Target identification and validation: This involves identifying a molecular target that is implicated in a disease process and validating its role in the disease. This may involve genetic, biochemical, or other methods.
- Lead discovery and optimization: Once a target is identified and validated, the next step is to identify potential compounds that can interact with the target and modify its activity. This involves using high-throughput screening and other computational and experimental methods to identify lead compounds. These leads are then optimized through medicinal chemistry to improve their potency, selectivity, and pharmacological properties.
- Preclinical testing: Once a lead compound has been identified and optimized, it undergoes preclinical testing in animals to assess its safety, pharmacokinetics, and efficacy.
- Clinical trials: If a lead compound passes preclinical testing, it can then proceed to clinical trials, which involve testing the compound in human subjects. Clinical trials are conducted in several phases and are designed to assess the safety and efficacy of the compound.
- Regulatory approval: If a compound passes clinical trials and is shown to be safe and effective, it can then be submitted to regulatory agencies for approval. The approval process involves extensive review of the safety and efficacy data and may take several years.
- Post-marketing surveillance: Once a drug is approved and on the market, it is subject to post-marketing surveillance to monitor its safety and efficacy in the general population.
The drug discovery process is a complex and expensive undertaking that requires collaboration between scientists, clinicians, and regulatory agencies. It can take many years and cost billions of dollars to bring a new drug to market.
What are the biggest advancements in drug discovery today?
There are several significant advancements in drug discovery today, including:
- Artificial intelligence and machine learning: These technologies are revolutionizing drug discovery by enabling faster and more accurate predictions of drug-target interactions, as well as the identification of novel drug targets.
- Precision medicine: This approach aims to develop targeted therapies that are tailored to an individual’s genetic makeup, lifestyle, and environment. It has the potential to improve the efficacy and safety of drugs by reducing adverse effects and increasing treatment response rates.
- Gene editing and gene therapy: These technologies are opening up new possibilities for treating genetic diseases by allowing scientists to modify or replace faulty genes. They have already shown promising results in clinical trials for diseases such as sickle cell anemia and inherited blindness.
- Novel drug delivery systems: These technologies are improving the efficacy and safety of existing drugs by enabling targeted delivery to specific tissues or cells, reducing side effects and increasing treatment efficacy.
- Open science and collaboration: The sharing of data, resources, and expertise is becoming increasingly important in drug discovery, as it allows researchers to work together to accelerate progress and reduce duplication of effort.
Overall, these advancements are helping to accelerate the drug discovery process and improve the chances of success in developing new and effective therapies for a wide range of diseases.
What is computer aided drug design?
Computer-aided drug design (CADD) is a computational approach to the drug discovery process that uses various computational techniques to design and optimize small molecule drugs. CADD allows researchers to screen large databases of potential drug candidates and simulate their interactions with target molecules or proteins, thereby reducing the time and cost of traditional experimental drug discovery methods.
CADD uses various computational techniques, including molecular modeling, molecular dynamics simulations, quantum mechanics, and artificial intelligence, to predict the activity, selectivity, and toxicity of small molecules. These techniques are used to optimize the drug candidates and to guide the selection of the most promising compounds for further development.
Molecular modeling techniques can be used to predict the three-dimensional structure of a target protein and to design small molecules that bind to the protein with high affinity and specificity. Molecular dynamics simulations can be used to simulate the motion and behavior of molecules in solution and to predict their binding interactions with target proteins.
Quantum mechanics can be used to study the electronic structure and properties of molecules, allowing researchers to predict their physicochemical properties and to design molecules with optimized pharmacokinetic properties. Artificial intelligence and machine learning can be used to predict the biological activity and toxicity of molecules, allowing researchers to prioritize the most promising compounds for further development.
Overall, CADD is a powerful tool for accelerating the drug discovery and development process and improving the success rate of drug development. It allows researchers to design and optimize small molecule drugs with greater precision and efficiency, ultimately leading to the development of new and effective therapies for a wide range of diseases.
Are Computer assisted drug design and Computer Aided drug design the same?
Yes, computer-aided drug design (CADD) and computer-assisted drug design (CADD) are essentially the same thing. Both terms refer to the use of computational techniques and tools to aid in the design and optimization of small molecule drugs. The two terms are often used interchangeably in the scientific literature.