The research and development culture established at Genentech by Art Levinson and Sue Desmond Hellman emphasized extreme scientific rigor and patient-centric view, values that inform the way I evaluate startups now as an investor. At Genentech, every drug discovery project was rigorously reviewed at the research review committee, and I remember the nerves I felt when I first presented one of our clinical programs to this incredible group of scientists and drug discovery leaders— they were not shy about their opinions, and questions came flying at you on your first slide. That intense scrutiny kept us honest with ourselves and on track with our development— increasing the probability of success and maximizing the value for patients.
Richard and I recently spent an hour together discussing today's startup landscape in science and medicine, and he compared Genentech's culture to what he sees as a concerning drop in scientific rigor inside some new companies. He attributes this to the intense pressure to publish, coupled with the increase in venture activity in life sciences. "One has to ask: how rigorous is life science really? Approximately half of the conclusions in papers and half of the data in publications is not reproducible. This gives rise to really difficult issues in industry. If you're going to spend tens of millions on a drug discovery program, you better make sure the data it's based on is correct."
This new generation of life sciences startups could benefit from emulating the Genentech review process. "You have to create a scientific culture that rewards people to get an answer, not the answer that keeps the project going. High-level review by the most rigorous people in the company who are not as emotionally attached to the programs results in better overall science."
After fourteen years at Genentech, Richard joined GV portfolio company 23andMe as Chief Scientific Officer and Head of Therapeutics, where he focused on translating genetic data into discovery and development of new drug therapies. "We had to develop a common language between the statisticians and the statistical geneticists in order to talk seriously about drug development," he says.
23andMe has the largest re-contactable research database of genotypic and phenotypic information in the world. The company is using its research platform to discover and develop novel therapies and has uncovered a number of promising targets. Richard adds, "There are many uses of human genetics that we're just beginning to experience. Imagine if you could develop a polygenic risk score in a particular disease and it was extremely likely that the people would show symptoms of the disease years faster than someone in the general population. Then enroll these people in clinical trials, and you would be able to see a signal much faster, with a much lower number of folks in the trial, and do drug discovery much more efficiently and much faster."
I've seen this work well in oncology with the advent of precision medicines. For Genentech's development of Herceptin, we only enrolled patients in clinical trials where we knew the target, and the drug only hit that target. This approach, however, doesn't easily map to other therapeutic areas. "It won't be as straightforward as oncology. It's going to be harder, slower, and more expensive," Richard adds.
For more complex, polygenic disorders, genetics provides only a partial clue about what might be important as a drug target. "You really have to understand much more about the biology," explains Richard. " Connection to the disease is somewhat less direct than an oncogenic driver. There's a lot more biology that needs to come into play, a lot more thought, perhaps a lot more preliminary experimentation around a therapeutic hypothesis. We're all looking for the next PCSK9, and I have news for you: There aren't very many."
With this in mind, I asked Richard how drug discovery will evolve. We've made tremendous progress in understanding the human genome and the roles specific genes play in various diseases. The next question is how these genes actually function. "That's a chemistry, biochemistry, even a physics problem," he says. "Life scientists are trying to take biology and make it a chemistry problem! And we've made pretty damn impressive progress on that. There's a terrific substrate to work with in starting companies nowadays."
We both spend a lot of time helping young companies get started, and as a GV advisor, Richard and I often team up to meet with new companies. When we evaluate potential investments, a founder's character is as important as the science. "You have to have a sense of practicality to go along with the vision," Richard says. "I like a down-to-earth visionary. Someone with a sense of humility. Arrogance is something that has always bothered me, and companies that start off being arrogant often frankly don't succeed."
Find more of my conversation with Richard here, where we discuss breakthroughs in the COVID-19 vaccine.