While 2020 has changed the way we live, the pandemic may also deliver lasting changes in how we work—and, as machine learning aims to have a larger impact on the life sciences, they may be changes for the better, according to insitro founder and CEO Daphne Koller, Ph.D.
In an interview for Fierce JPM Week with Amirah Al Idrus, Koller laid out how the industry’s embrace of digital tools to keep research moving in the face of COVID-19—from video conferencing with potential partners and investors to distributed clinical trials that equip patients at home with remote devices—could be adopted by biopharma companies for the long haul instead of simply as short-term solutions.
At the same time, these technologies provide an opportunity to deliver the massive amounts of bespoke, fit-for-purpose data artificial intelligence programs have been hungry for.
“Machine learning was transforming the world sector after sector, but it really wasn’t having as much of an influence in the life sciences,” Koller said. “The critical limitation was the availability of high-quality data.”
“I realized that now we’re finally at the moment in time where machine learning has matured to the point it could make a difference, and that biology had developed a suite of tools that enable the right kinds of data to be generated at scale,” she added.
RELATED: Insitro nets $143M to ramp up machine learning drug discovery
This can include broader information that better captures a potential treatment’s effect. For example, in muscular dystrophy and neurodegenerative disorders, observing a study participant’s day-to-day life may provide a clearer picture of their status over time, instead of having them come into a clinic to perform a one-time test.
Those physical tests and subjective observations may not fully reflect the therapy’s impact in the real world—besides, people naturally try harder to complete tests because they want to do well, which can further muddy the data, said Koller, whose insitro began working with Bristol Myers Squibb late last year in amyotrophic lateral sclerosis and dementia.
“People are still recording stuff in little notebooks, which are transcribed by a nurse, and sometimes are actually faxed over,” she said. “When you think about that, it’s medieval, right? And I think that’s going to go away with this new world of distributed clinical trials.”
Remote technologies can also help make research studies more inclusive by tapping into a broader patient pool less limited by geography.
RELATED: Koller's insitro snaps up Haystack to build up machine learning tech
The same concepts apply to the workforce: With remote, distributed employees, companies can recruit more talent “and also let people live in more affordable areas than just the biotech hubs of Boston and San Francisco,” Koller said.
At the same time, shepherding major transactions through virtual meetings—as insitro’s $143 million series B round was completed in May, during what was considered the peak of the pandemic at the time—will become more efficient without having to crisscross the country to meet face-to-face with investors.
“You can really have an ongoing series of conversations, and really dig into the science in a way that, if you’re making a trip once, you’re not able to do,” she said. “I think that will shift deal-making, both on the partnering and on the investment side in ways that will not change once the pandemic is over.”
For 2021 and beyond, insitro will look to build out its own pipeline of potential therapies, with “eminently prosecutable” targets in neuroscience and liver diseases borne out of the company’s previous work on human genetics and modeling cellular systems.
“We hope by the end of 2021, we can develop chemical matter against some of those targets and enter 2022 with an actual pipeline that will help us get drugs to patients.”