Five Questions With a16z’s Vijay Pande on AI and Making New Drugs

(Page 2 of 2)

Evan’s work has demonstrated this not just in potency, or how well a drug binds to its target, but also in ADME [absorption distribution metabolism excretion] and toxicity [studies]. And these are areas where I think there’s naturally going to be a huge win in making advances toward avoiding issues that you’d find in Phase 1 clinical trials.

X: Do you anticipate competition in the AI/drug design sector will thin out once the strongest companies pull ahead, or that many businesses will be able to thrive in the space?

VP: In some ways, AI is like biochemistry; all of pharma uses biochemistry right now, and molecular biology, so it’s not that just one company we’d rely on to … It’s now become a key tool. Our thesis is that ML and AI will become prevalent in drug design in an analogous fashion. One of the key differences, though, is that if you view this just purely as a technology, you probably miss out. The big win here is realizing there is this opportunity to shift from discovery to design, from science to engineering, (and) from serendipity to planning, and that shift is something that ML really facilitates in a way that I think other technologies haven’t. A lot of what we’re looking for is not just a strong technology, but that engineering mindset as being prevalent within the company, and I think this is something that is going to be much more common in drug design companies.

X: Why is this mindset important for today’s biotechs, especially those deploying AL/ML for drug design?

VP: The benefit of an engineering approach here is that these companies are building platforms that can identify multiple targets that people never have found before and hit those targets much more rapidly. Now the interesting thing is that part of the shift to the engineering mindset is a shift to portfolio mindset. You’re not just going to have your one compound, your single-asset company where that’s your baby and the whole company lives or dies on a readout in trials, and it either goes to a few billion or to zero. It’s more about: let’s build a portfolio, and even use portfolio theory, to think about how can we best mitigate risk, just like an investor would.

Single PageCurrently on Page: 1 2 previous page