While training in the UCLA M.D./Ph.D. program in neuroscience, Alice Zhang made a keen decision to launch a genomics startup that would accelerate and lower the cost of the drug discovery process by combining innovation in neuroscience, machine learning, and genomics.
That company, which she launched in 2015, is Verge Genomics, whose vision is to rapidly develop lifesaving treatments for some of the biggest unaddressed diseases of our time – Alzheimer’s disease, ALS, and Parkinson’s disease. It currently takes about 12 years and $2.6 billion to get a single drug to market. Verge Genomics is trying to address these problems by making drug discovery faster and cheaper.
In just a couple of years, Verge has assembled a group of leading experts in machine learning, neuroscience, drug development, applied math, biophysics, and statistics from UCLA, Stanford, Oxford, and UCSF. Verge also invested heavily in the creation of its own proprietary datasets, generating one of the field’s largest and most comprehensive databases of ALS and Parkinson’s disease patient genomic data through partnerships with a dozen top-tier academic and government organizations.
The Silicon Valley startup recently raised $32 million in Series A financing, led by DFJ. New biotech investors WuXi AppTec’s Corporate Venture Fund, ALS Investment Fund, Agent Capital, and OS Fund also participated in the round. The investment will help Verge advance its ALS and Parkinson’s disease drugs. The funding will also be used to expand the number of diseases Verge has in its portfolio.
Zhang, Verge’s CEO, has been at the forefront of systems biology research for over seven years at the National Cancer Institute, the Lewis-Sigler Institute for Integrative Genomics at Princeton University, and UCLA. During this time, she co-authored four peer-reviewed papers in high profile journals such as Cell and Neuron. She graduated from Princeton University magna cum laude with high honors in Molecular Biology and is a 2012 recipient of the prestigious Paul & Daisy Soros Fellowship for New Americans. Zhang was recently named as a Forbes 2017 30 Under 30 featured honoree.
WuXi AppTec Communications recently caught up with Zhang, who told us why machine learning and artificial intelligence can fast-track treatments for the world’s biggest unmet medical need – neurodegeneration.
WuXi: Can you give us some background on how Verge Genomics started and where the company is today?
Alice Zhang: At Verge Genomics, we’re using machine learning to develop new drugs more cheaply and quickly. Before I started Verge Genomics, I was an M.D., Ph.D. student at UCLA when I realized that drug discovery is broken. Right now, drug discovery is still a guessing game and it takes more than a decade and billions of dollars just to develop a single new drug. But at the same time, we’re in the middle of this exciting convergence where advances in machine learning and genomics has created this unique opportunity to take the guesswork out of drug discovery. Diseases that we previously thought were incurable, such as Alzheimer’s disease, are within reach. I was so impassioned about this that I left my program three months before graduation to start the company. I founded Verge with the vision of building the world’s first pharmaceutical company that uses machine learning to tackle neurodegenerative diseases with artificial intelligence. We are a full stacked drug company, which means that we have both the front end of the proprietary machine learning algorithms and our own back end in-house drug discovery and animal labs. In the last year, we’ve been able to develop several promising drugs that are currently in development for ALS and Parkinson’s disease. We’re currently a team of 14 people that combines some of the top talent in machine learning and mathematics who sit side-by-side with neuroscientists and pharma veterans.
WuXi: You said Verge Genomics tends to discover drugs faster and cheaper. How fast compared to traditional pharma companies?
Alice Zhang: We’ve been able to identify promising compounds within the last year and a half, where it takes most pharma companies multiple years to get to the same point. We’ve only needed to test on the order of 70 drugs, before we first started identifying ones that were working, instead of hundreds of thousands of drugs.
WuXi: Many startups tackle cancers first because it’s a hot area and more likely to have success. Why did you decide to tackle CNS diseases such as Alzheimer’s first?
Alice Zhang: We tackle neurodegenerative diseases for two reasons. The first is that neurodegeneration is the world’s biggest unmet medical need. Alzheimer’s disease, for example, is the one of the only diseases with growing death rates over the last decade. Yet, there are no existing drugs that stop or even slow the progression of these diseases. One major reason for failure is that our existing models of these disease are too simple to address the enormous complexity. Where machine learning can make the biggest difference is in the most complex diseases, where we can have the biggest competitive advantage over a traditional pharma company.
WuXi: Can you describe your platform at Verge? And how could it potentially help find the answer to some of these diseases?
Alice Zhang: In the last couple years, we have seen this really exciting growth of new machine learning technologies and an explosion of genomic data. Over the last decade, the cost of genome sequencing has dropped even faster than the cost of computing. We’re in the middle of this unique time where we finally have the tools to tackle these diseases . Specifically, at Verge, we use all genomic data to improve on the current drug discovery model in three primary ways. The first is that instead of looking at one gene at a time, we use our algorithms on human genomic data to map out hundreds of genes that are causing disease. Secondly, instead of relying on animal data until clinical trials we use human data from day one to identify the relevant genes to understanding disease. Lastly, instead of blindly screening drugs, we computationally predict genes causing disease, and then we use that to predict a drug that will targets the disease genes. So, instead of just screening hundreds of thousands of drugs, we can only test 70 before we actually start seeing drugs that are starting to work.
WuXi: How is the data used on your platform? How does it differ from other platforms?
Alice Zhang: One big difference at Verge is that we generate a lot of our own proprietary data. Unlike many companies that exist right now, which usually rely on public data, we generate our own data by collecting thousands of patient brain samples and spinal cords and we sequence them internally ourselves. So that has given us enough data; one of the biggest values of the Verge platform is we have a huge amount of patient data, and then the second important point is that we don’t just make predictions and leave them hanging. We have our own internal labs and animal facilities that generate validation data. So, every time we make a prediction we quickly test that in the lab, validate that hypothesis, and generate a huge amount of data that feeds back into the platform which continues to improve it and retrain it. We also have proprietary validation data, and those two combined is what feeds the machine learning algorithms and helps them improve over time.
WuXi: We recently read in a paper that they used fixed data to try to find the true roots of Alzheimer’s disease. They argued that infection might contribute to the development of the disease. What do you think about that, and do you think your platform will give us new insights into the disease?
Alice Zhang: Yes, in fact we have already done that. Often times, investigators will study individual hypotheses about what causes disease in insolation. For example, some people might think that it’s an infection. Other people might think that is a defect in protein processing. What is nice about taking an unbiased system’s approach is that instead of letting human bias influence what we pursue in studies we can actually have a model that integrates multiple mechanisms and takes into consideration the whole landscape and a snapshot of what disease looks like. We’ve also shown that we can find and validate new mechanism for disease; we published a paper in Nature Medicine in February with our collaborators that identified the new mechanism of disease.
WuXi: The industry has a lot of hope for AI in drug discovery. But we still don’t have an approved drug using AI. Do you think AI in drug discovery has a real hope or is it just hype?
Alice Zhang: There’s a lot of excitement in the field. I think there are real trends that are indicating that there’s going to be a change in the field, the biggest of which is the improvement of data availability. I think there have been advances in genomic sequencing and the cost of genome sequencing that have generated vast amounts of data sets. This makes it finally possible to start running machine learning. That being said, I think that there are different companies that are making different bets about the application of machine learning. There are some applications that are more promising than others. One of our core thesis at Verge is that to realize the full potential of AI we need to integrate computation and biology. We can’t have it exist in these silos anymore. That’s why it’s so important for us, from day one, to have an integrated team. Right now the team is 14 and that is not only machine learning and mathematicians, but they sit right next to neurobiologists and pharma veterans; we don’t just have a computational platform, but we also have our own lab and facilities and that’s so important because in the field of AI, it doesn’t matter how sophisticated your algorithm is if you don’t have the sufficient data from which it learns. So that’s why it’s so important to have our own labs so that we can actually generate that validation data that feeds into the algorithms and improve it over time. I think the most promising uses of AI in biology in drug discovery are from those companies that actually have both sides of the coin where they have not only the computational horsepower, but they also have the biologists on the team and the labs to actually generate that close up feedback loop.
WuXi: Can you elaborate more on your ‘unbiased’ software?
Alice Zhang: The challenge with these diseases is they are oftentimes very complex and there are multiple mechanisms that are causing the disease. So, we look at the general landscape of how genes are behaving across all 20,000 genes that are in our genome. We look for sets of genes that are always defining in Alzheimer’s brains but not in healthy brains and ask how can we identify drugs that turn off that entire set of genes as a whole to make the gene expression landscape look more like a healthy brain? It’s really about taking a data driven approach to finding drugs.
WuXi: What do you envision Verge will look like in the next one-to-two years? And what is the next milestone for the company?
Alice Zhang: Our focus is first advancing our current drugs along closer to the clinic Also expanding our team so that we can start bringing on chemistry – so hiring chemists as well as translational medicine scientists. We’re also expanding our pipeline so we’re adding more diseases to our portfolio.
WuXi: Since the recent investment news, should we expect a stronger collaboration between WuXi and Verge?
Alice Zhang: Yes, WuXi is a major investor in Verge and we’re excited to be working with WuXi. Right now, we’re exploring potential leads to collaborate. WuXi is a big resource for our medicinal chemistry needs.
WuXi: Can you describe Verge’s business model?
Alice Zhang: We’re building a pharmaceutical company; we’re developing the drugs themselves and we benefit directly from the decrease in cost and the time it takes to get there. That’s what’s made so many of our investors so excited to invest in the company is that there is huge upside for actually owning the drug product that you’re developing. Our long-term hope is that if we prove that our first couple of drugs in human proof-of-concept, the entire industry will start to realize this is an approach that needs to be used more broadly. Specifically, that we need to be using human genomics to study disease, that these diseases are causing multiple genes not just single genes, and not rely so heavily on just animal models moving instead to a more holistic viewpoint of the disease.
WuXi: Verge already has a robust pipeline. What other disease areas are you looking at?
Alice Zhang: We are also working on Parkinson’s disease. We’ve identified a dozen targets in Parkinson’s disease that we’re developing right now.
WuXi: Let’s talk a little bit about you and your career. Why did you stop your studies early and start Verge? Was there an urgency to do this?
Alice Zhang: I think it’s such a unique time right now – we’re going through a revolution where the genomics and the neuroscience are converging. And I think it’s a matter of who gets there first. I started Verge when I did because Verge has become one of the first movers in combining neuroscience and genomics through drug development, and that first mover advantage is very important in biotech. As I was getting my Ph.D., I was writing a publication on my work and I had this thought, do I just want this to be published and just sit on a bookshelf somewhere or do I want it to actually be commercialized and to help patients? I also realized that if I wasn’t going to be the one doing that, there are actually very few people in the world that are as well equipped to do that. At the time, we applied to Y Combinator and we got in for that summer and so that’s when I decided to leave school to start the company.
WuXi: Do you think in the future a Ph.D. degree will be a requirement for young people who would like to achieve success in setting up a biotech company?
Alice Zhang: My Ph.D. experience was very valuable. I couldn’t have started a company without the actual experience I got in becoming an expert in my field, so I think that part of it is valuable. I think having a Ph.D. can still be useful, but in terms of starting companies, I think that we’re going to start seeing a trend where there is going to be more and more scientists that start companies that don’t look like traditional biotech companies. I think is there’s going to be a huge growing influence of technology in biotech. You’re going to have these younger and savvier teams that are truly understand artificial intelligence, who grew up with genomics. These same companies are also going to take less money to start. Whenever you take less money there’s a lower barrier to entry to starting the company and that’s when you start getting a democratization of company formation. I think it’s a similar trend to what you saw happen in technology in the ‘90s and 2000s where the costs of starting a company just went down and the number of younger CEOs went up.
WuXi: What advice would you give young people who are thinking about entering this field?
Alice Zhang: My advice for them is first to learn how to articulate and communicate your science; first understand why your science is valuable and to whom it is valuable, whether that’s patients or other companies. I think that once you can do that, then you can really do more than you think – and not necessarily within traditional constraints. Don’t be afraid to pursue that if you see an opportunity.