This Startup Has Raised Nearly $100 Million To Make Good On The Promise Of AI For Drug Discovery

David Berry

Flagship Pioneering

^For many entrepreneurs, founding a new company can be a momentous occasion, often being the culmination of a lot of time and effort. For David Berry, it’s a Tuesday. Since joining VC firm Flagship Pioneering in 2005, the 42-year-old has founded or cofounded 30 different life sciences and tech companies, leading over 10 of them as founding CEO. 

His latest company, Valo Health, is one of many aimed at using the latest advances in AI, machine learning and computation to improve the notoriously difficult and inefficient process of drug discovery.

“We’ve been operating with a drug development model for years that has a low probability of success,” he says. He’s got a point - fewer than 10% of promising drugs that begin the clinical process end up making it to market, and that figure has remained constant for years. “I see an opportunity where data and compute give us a fundamental opportunity to rewrite the drug discovery process.”

Berry has a lot of firepower behind his vision. The company was founded around 18 months ago and in July 2019 raised a series A round of nearly $100 million from Flagship Pioneering and PSP Investments. In November 2019, Valo acquired San Francisco AI company Numerate, folding that company’s operations into its own. In March 2020, it acquired FORMA Therapeutics’ early-stage discovery platform, which included its team, IP portfolio, two early discovery labs and R&D libraries. The initial focus of the company will be on finding small molecules that can be used to treat cancer and neurodegenerative diseases. 

“The whole thought process of reimaging this entire drug development cycle was very, very exciting to me,” says Ron Hovsepian, an executive partner at Flagship. “Being able to use all this exciting stuff we’ve done around cloud and data to create a whole new drug discovery process is really important.”

Valo is far from the only company focused on using AI to improve the drug discovery process - there are over 200 startups in the space, such as Insitro, founded by Coursera cofounder Daphne Koller. It’s a trend that’s going to be increasing as time goes on, in part because of workflow changes due to the Covid-19 pandemic, says Michael Shanler, an analyst at Gartner. “I always see questions about how AI can be leveraged,” he says. “So the trend is real, but I’m not convinced most of the companies are very differentiated.”

“What we do with that data is let biology reveal itself as to what’s going on.”

^Over a conversation on Zoom, Berry sits in a room overflowing with books and the question of what sets his company apart doesn’t faze him. If anything, it gets him more excited. First, he says, Valo is focused on “high density longitudinal data” from thousands of sources containing information about how different diseases and treatments affect people, including data of what those people look like when they’re healthy. This way, he says, his researchers can get a better understanding of disease states. “What we do with that data is let biology reveal itself as to what’s going on,” he says. 

The second differentiator, he says, is Valo’s computational platform, Opal, which it uses to develop models that predict what types of molecules might make for useful drugs in a fully integrated approach that looks at not just a particular biological pathway, but the whole patient. “Yes, we are finding new targets,” he explains, but his company is looking beyond that. “What are the features associated with that biology to get therapeutics to be successful. What is the clinical trial? What is the patient population?”  

By taking a broader view, Berry aims to “take into account the complexity of what goes into drug discovery and development.” One consequence of this is that the Opal platform can be used across multiple points of drug discovery. Valo is interested in new targets and developing its own drugs, but Berry says the platform could be offered as a service to other companies as well. “We do see opportunities to partner across the value chain,” he says.

“Hard problems and easy problems are both equally difficult. So why not focus on hard problems?”

^Raised in Mt. Kisco, New York, Berry did his undergraduate work at MIT, receiving a Bachelor’s in brain and cognitive sciences. Five years later, he was awarded a PhD in biological engineering from MIT and received his MD from the Harvard-MIT Health Sciences and Technology program a year later. The research he undertook in school looking at innovation in cancer and stroke treatments earned him a Lemelson-MIT student prize in 2005. 

Since joining Flagship Pioneering in 2005, where he’s now a general partner, Berry has been involved in founding over 30 companies, including microbiome company Seres Therapeutics, biotech company Axcella, green chemical company LS9 (which was acquired in 2014 by Renewable Energy Group for $61 million) and agtech company Indigo Agriculture. 

It’s to this extensive experience that Berry credits the passion he has for Valo. “I looked at the learnings from several companies I’ve founded over the past decade and a half and saw that I could leverage that experience to build a bigger company, to build a broader company, to build a company that has the potential for systematic impact.”

^Over the past 18 months, fueled by the Series A, the company has grown significantly. It already has over 100 employees, in part from the two acquisitions. “We’ve been able to harness great work that others have been doing,” says Berry. 

Numerate, which Valo acquired in November 2019, was a Bay Area-company founded in 2007 that was focused on a computational approach to drug discovery. Its founder & CEO, Guido Lanza, now serves as Valo’s head of innovation. Numerate was focused on using its computational models to screen promising drug candidates, with an eye to making sure the best moved on in the development process. 

Lanza’s years in the industry, however, led to frustration with the inefficient ecosystem of drug development—a frustration he bonded over with Berry on their first phone call. The two clicked quickly. “We went from a phone call to a meeting to a term sheet in a couple of months,” recalls Lanza. 

Numerate’s computational models combined with the Opal platform Valo developed were enhanced further once combined with the assets Valo acquired from Forma Therapeutics, Berry says. “That allowed us to create a truly closed loop cycle for our discovery pipeline that reduces time, reduces costs and increases the chances of success.”

Berry fully admits that his company is taking on some very hard problems in a big way. That just seems to excite him more. “When I was getting my PhD,” he says. “[Biologist] Bob Langer said something really interesting to me, which is: ‘Hard problems and easy problems are both equally difficult. So why not focus on hard problems? Because when you solve them, you make more of a difference.’ That’s made a lasting impact on me. ”

#News | https://sciencespies.com/news/this-startup-has-raised-nearly-100-million-to-make-good-on-the-promise-of-ai-for-drug-discovery/