NewHavenBIZ

12 n e w h a v e n B I Z | N o v e m b e r / D e c e m b e r 2 0 2 0 | n e w h a v e n b i z . c o m A r t i f i c i a l I n t e l l i g e n c e Big Data CT bioscience companies use AI to speed up life-saving drug discoveries N ew Haven's BioXcel erapeutics has been a company for only three years, yet it's on track to have its first commercial drug by the end of 2021 — a dissolving film that's similar to a mouthwash strip and can quickly calm agitated patients. By Natalie Missakian "We went from first-in-human [dosing] to completion of the phase 3 trial in a 19-month time period, which is unheard of in our indus- try," says CEO Vimal Mehta, who announced the drug's successful phase 3 trial results in July. In fact, industry experts say it typically takes 10 to 12 years, and as much as $2.6 billion, to bring a new drug from lab to market. And that's only if that drug is one of the few that make it all the way through clinical trials. But BioXcel is one of a growing number of drug companies trying to slash that timeline by more than half — and increase the return on their R&D investment — by using artificial intelligence and machine learning to assist with drug development. e company, which spun off from data analytics firm BioXcel Corp. in 2017, uses proprietary machine learning algorithms to si through tens of millions of pages of scientific research hunting for older drugs that have been overlooked or discarded, but can be repurposed to treat new diseases. Mehta says there are more than 5,000 new research papers published in medical journals every day, while one study shows the average scientist can read only 314 in an entire year. "Even if you're a Nobel laureate you can't keep all of that information in your head. It's humanly impossible," he says. "With a machine we can grind 40 million to 50 million publications in a few hours, literally." By focusing on drugs that have already been found safe in human testing, BioXcel saves time by jumping directly into later-stage clinical trials. Its agitation drug is a reformulated version of a 20-year- old Pfizer drug, marketed under the brand name Precedex, that had previously been used as an intravenous sedative in the operating room. Growing trend Nationally, the number of AI-focused biotechs is growing. In 2014, there were only 30 U.S. biotech startups claiming to use sophisticated AI platforms for drug research and development; today there are 150, according to Deep Knowledge Analytics, which has studied the trend. Meanwhile, more than 40 BioXcel Therapeutics CEO Vimal Mehta says the company used artificial intelligence to speed development of its drug for agitation, which is on track for FDA approval in 2021. PHOTO | STEVE LASCHEVER

• Cover