Name: Drug Development in the Age of A.I.
Uploaded: 2019-04-03
Duration: 9 min 52 s
Description: As we transform drug development in the age of A.I., fine-tuning our patient understanding will be more important than ever.

We've known each other, you know, for a long time. And you have the wildest career. You've gone from one of the great biotech companies in the world, you know acquired by a big pharma company, worked at Google, and now head of r & d GSK, it's a pretty wild transition between all of. So, ending up at Big Pharma's, is that where you thought you would be? Yeah. Why? Well, I think that once I left Academia at UCFF as a Cardiologist, I wanted to see how I could impact as many patients as possible and while there's a lot of exciting things that happen in biotech, some of the most innovative things I think the scope and the scale that you can do things in big pharma when it's going well, when everybody's rowing in the same direction the impact that you can have is pretty profound. So I'd always hoped to have a role like this to head up R & D end to end, to be able to impact patients' lives in a meaningful way this way. You're unique, in that, they let you stay in San Fransisco and yet this is basically a UK based company, UK, Philadelphia and your in San Fransisco. The pipeline you're kind of building, almost from scratch. And your building a new mentality in how to approach drug developement. So tell us a little bit about, the idea of, you know, your ideas of functiongenomics and others to take that heat rate, from a relatively low number that they big pharma complaints all the time. We have to charge big prices because we are not officially developing drugs and your trying to flip that a little bit. Yeah, you know, when I came to Jeski about 14 months ago, it has a terrific pipeline. Actually, there's a lot of leaders in HIV leaders and respiratory for years. A very innovative immunology portfolio actually about 30 or so drugs in the clinic and immunology, but what's plaguing GSK and I think all farmers you said is that, unfortunately, want to drug enters clinical testing 90 % of the time it fails to reach the end game to help a patient. This valuated 90 % is just an abstrast for the entire industry and so we spend a lot of time asking what's teh biggest impact we can have in terms of improving that and what we arrived at is based on some work actually ideas came from some sceientist that shows that if you can find a target that genetically validated by that I mean that That you have evidence from human genetic studies that show, that when you have this genetic abnormality, you get a disease. And we can isolate the protein that's causing that, and then transform it back to its normal type or block it if it's causing the disease, that you're twice as likely to have a drug. And that going from 10 to 20, actually is transformation for the industry, can think about developing twice as many drugs for the same price or being able to Develop, you know, a drug for half as much you want to look at the math. But we think that that's an exciting opportunity and with the explosion of the genetic data that's available for people, for patients throughout the globe and the cost coming down dramatically. And we've actually done three or four different collaborations with UK Biobank with open targets with Finjan and most recently 23andme. We think we can tap into this enormous amount of data that exists to find novel important targets that hopefully will turn into medicines. That's kind of novel, I mean, the head of r & d at one of the biggest pharmaceutical companies buying data from a consumer genomics company. Yeah, it's very novel. I think when you think about the mission of [ UNKNOWN ] was really to some day have a big impact on patients. And I think we both realized that the impact we're trying to have is better targets, they're trying to have a big impact on the people that signed up for their service, those who were interested. And we really felt like it's one of the great examples of an absolute win- win on both sides. And it's Wonderful company it's the great culture. You mentioned San Francisco. It's Richard Scheller who is running the 23andMe Discovery and I were heads of our RND. Respectively at Genentech's. We've known each other for a while and I've known Ann for a while. It was perfect cultures, perfect opportunity. I think we're gonna develop a lot of great drugs from that. so the naysayers will say listen this is self- annotated data, I think I have this disease and that disease. How do you get over that self annotation and get to something that you believe in Well there's three or four ways, first of all and we've learned this lesson from google size matters a lot and when you have the kind of sample size they have there in the millions I won't say exactly how many but But it's many millions of patients people are enrolled in the service. And there's been some studies they've done internally, when they ask a person in the, in the 23andme database. Do you have psoriasis as an example, or something like that. You can go back and look to see if those patients have the known genetic associations that have already been published in the literature. And if you find the 7 or 8 different GWAS findings that were already reported, you have a good sense that the other 10 that are new are probably real or at least most of them are. So this sort of validation through the published literature is one way. But we also have collaborations with UK bowel bank where it's much smaller it's very deeply phenotyped and we can use that with open targets and [ UNKNOWN ] and other Genetic data set to validate something, so I think this is a very powerful thing. As you also say, many of the targets are brand new and we're really introducing functional genomics, which you are very familiar with, where we can actually interogate cellular level. What is happening and see if we can get biologic validation of these associations to understand. What is that base pair change. You got 3 billion base pairs, one changes and you get a diseases. What's going o at the cell level and then biologically that makes sense and we can [ UNKNOWN ] maybe in [ UNKNOWN ] models and maybe validate through other drugs that have been developed et cetera. It gives you more confidence, so I think it's a triangulation of lots of data points. So you stol me something back stage somethingwhich astounded me Is that when you look at a disease like ovarian cancer that has a particular mutation that says you'll respond to a drug, BRCA, and PARP inhibitors. And there's an entire cohort of patient with functional [ UNKNOWN ] where you can show that they have the same phenotype Yes, they don't have that [ UNKNOWN ]. Tell us a little bit about that day because it's an amazing day. Well, one of the thing that we really focussed on is this funtional genomics. What's going on at that. Not just the structural level, that's your base pairs. What are your 3 billions base pairs? 23 and me can tell you. You can do all the sequencing. What is the function? What is functionally happen? So not just is our mutation but it's the gene working. The gene is the network. How is it all functioning? And so one of the interesting observations and one of the drivers for us looking at a company called the sorrow and, ultimately, acquiring them was that the synthetically family concept where you can have a PARP inhibitor that doesn't do much bracket mutation that doesn't do but the two together a profound impact. And we asked a question that sorrow folks in the global community aspect. Are there. Like other defects and through this functional genomics assessment, you can see many other proteins like a try at home and read 50 and the thing Coney proteins, etc. that all seem to have this homologous recombination defect the similar kind of Brock like, and in clinical trials that looks like the PARP inhibitor that we have proper of is working, not just in the 15 % of ovarian cancer patients that have the. The BRCA mutation, but in the patients who have this functional BRCA alike, so we have a big trial coming out at the end of the year, that we are optimistic will help identify or maybe as many as three times as many patients who could benefit from this therapy. So, it's not just developing a drug but it's actually figuring how to use it in the global sense that's critical. Well I've always said great drugs are discovered but transformational drugs are developed. And you have to find the other patients that might not be obvious, and how do you personalise them. That's not quite the end on one's but it's really amassing a data set that we can help the most number of patients to clinical trials. So I have to ask you this, I mean coz I think everybody wants to know is that, you spent time at Google looking at things like longevity. What did you learn? What was your take home from that experience? I mean it's a quite a concept that a search company got into this space at all. Let's see, well, one of the things that I confirmed was that, the person I got to work with Art Livingstone was one of the most terrific people to work with and a brilliant person and this gonna Continue to change the world. So that was one thing I learned to Calico. It Well, I guess the big picture and Calico that that I took away with is that as we age, I think of this model as a homeostatic reserve. We Rob basically buffering from disease is we're young, we have lots of buffer, we get insults and It's at each insult doesn't seem to cause disease because we're buffered. We're resilient. We're resilent. But as you get older, I kind of have the model that you become for certain diseases more or less resilient, but over time, your buffer capacity. And so that when we finally see in evidence of cliinical disease, it's probably not just the disease happened last week, and you became sympomatic. But that was probably accumulating and you've tipped over your bugger capacity. And that's sometimes early intervention is really gonna be the way to make the biggest impact on a lot of sort of age- related diseases in my opinion. So childhood cancer is predominantly curable. Yet once you turn 25, the same cancers become incurable. Does that go along with that concept? Maybe. I think, one of the things that I'm most facinated about another place is immunology, and the immune system changes quite dramatically as you age, and I think how all this is gonna play out is unclear, but, clearly, the immune system becomes less robust as you get older And that might be part of this homeostatic preserve that when you're young you have. We're also very focused on cell therapy, turning cells into medicines, and maybe we can rejuvenate these cells in a way that can make them more youthful and be able to cure some of these cancers that were in early childhood more likely to be cured. I think in our short conversation the audience certainly has a sense and I think a hope is that you're bringing a new approach and a new way of thinking to drug development. And it gives all of us some optimism that going forward, we're gonna have some shots on goal that we've never had before. So thank you for what we're doing and it was a true privilege. We appreciate it. Thank you Dean. [ APPLAUSE ] [ BLANK_AUDIO ].

Drug Development in the Age of A.I.

Tech

As we transform drug development in the age of A.I., fine-tuning our patient understanding will be more important than ever. 