In startup world these days, the word “biotech” is increasingly accompanied by “computational” and two, two-letter initialisms: AI and ML.
Those tools—artificial intelligence and machine learning, respectively—have been around for decades, but in recent years have become faster and cheaper, accelerating their use by those in the business of discovering and developing new drugs. Another startup looking to take advantage of those improvements, South San Francisco-based Genesis Therapeutics, has scored $4.1 million in seed funding and publicly joined the growing fray of biotechs with grand ambitions of disrupting the slow, costly process of discovering and developing new medicines.
Andreessen Horowitz, also known as a16z, led its seed round, one of a handful of seed-stage investments it has made in biotech. Felicis Ventures, another VC firm based in Silicon Valley, also invested. Genesis says it plans to focus on developing small molecule drug candidate for patients with “severe and debilitating disorders,” and that it aims to move ahead investigational drugs it discovers itself and in partnership with pharma.
 |
Want more AI, machine learning, and drug discovery content? Network and engage with industry leaders online this June. Learn more. |
The technology that underpins the company was invented in the Stanford University lab of a16z general partner Vijay Pande, who joined the firm in 2015 to lead its debut $200 million biotech fund. Since then the firm has invested in AI drug development outfits including Erasca, Insitro, and TwoXar, and raised $450 million for a second bio fund.
Pande recently talked with Xconomy about how AI will impact drug development, what differentiates Genesis, and why biotechs need to adopt a “portfolio mindset.” The conversation has been lightly edited and condensed for clarity.
Xconomy: What sets Genesis apart from the many AI drug development startups operating today?
Vijay Pande: This is technology that came out of my Stanford lab that I was running before I left to found the bio fund at Andreessen Horowitz, so I’ve known [founder] Evan Feinberg for five to six years, and I know the technology very well, so it was very natural for me to get excited about that part. The part I think really differentiated Evan’s approach here was getting a really great drug hunter like [acting Chief Scientific Officer] Dr. [Peppi] Prasit involved very early. I think that he’s often thought of as a “drug hunter’s drug hunter,” and Evan getting him on board I think is a huge win for filling that team and also a validation for the significance of that technology.
X: What is different about the software tools that Genesis plans to use to search for new drugs than the algorithms used by other such biotech startups?
VP: There [are] 200, 250 companies now in this AI/drug design space. Given the prevalence of tools like [open source ML framework] TensorFlow, algorithms in the public domain, and public data, it doesn’t take much to build something just with those off-the-shelf pieces that looks pretty good, especially compared to what people could do before. All of those companies, if they’re using basically the same algorithms, the same tools, and the same data, they’re going to get the same answers as each other. So differentiation is really going to be key …
Evan hasn’t just done what most people do, which is take algorithms that people use in computer vision, from identifying cats on the internet and that type of thing. … For images, it’s very clear what the [statistical] representation are: pixels. For molecules, it’s not clear at all. … One of the key advances that Evan and Genesis made is in that area of representation—how to think about the right way to explain what the molecule is to a computer. … They have figured out the right way to represent molecules such that AI and other algorithms can take advantage of that representation.
X: Genesis is a very early-stage company, especially compared to others a16z has backed in the space. What have these advanced algorithms allowed it to do that the firm believes will allow it to develop new drugs more efficiently than others?
VP: Five years ago, people nearly had to come up with the right features by hand. In a sense, their brains were the first part of the neural network. … With deep learning right now, I think the big difference is that if you have the right representation, deep learning can learn the right features from there.
