Brendan Frey, the founder and chief executive of Deep Genomics, thinks the wrong people are developing new drugs to treat genetic diseases. In fact, he doesn’t think humans should be doing most of the work.
It’s Frey’s view that the next wave of medicines will be discovered using artificial intelligence, and he has support from big investors to create a company that could do that: Deep Genomics on Wednesday announced a $180 million funding round led by Softbank Vision Fund 2, with participation from Fidelity Management & Research Co. and others, bringing its total venture haul to $240 million.
Frey, a Toronto-based pioneer in artificial intelligence research, saw in the early 2000s that researchers were unlocking vast amounts of data about the human genome, but biologists were not able to efficiently use it to make new drugs. He thought the process of analyzing the data — using it to discover drug targets, understand gene mutations, and predict outcomes — would be better suited for artificial intelligence, a computerized system that would improve over time.
“To make a successful drug, you have to solve complex biology, be accurate, and you have to be able to do things at scale,” he said. “AI is the best technology we have for doing [this].”
After more than a decade of research, Frey started Deep Genomics in 2015, realizing his systems had reached a point where they worked better than humans in a lab. Scientists at Deep Genomics have access to what he calls an “AI Workbench,” a system of 40 different AI-powered tools that can be used to identify the best therapeutic approach for genetic diseases.
“You just push a button on a computer,” he said. “It changes drug discovery from an act of serendipity . . . a ‘let’s try this out’ . . . to one of ‘let’s try everything out’ . . . every gene in the genome, every possible mutation, every drug we can throw at it, and then accurately quantify what the consequences will be.”
The goal is to save drugmakers time and money during research and development, and to give them access to information and patterns that humans alone are not capable of observing.
Deep Genomics is based in Canada, where Frey did his research, but he said the company is strategically “invading” Boston to engrain itself in the life sciences cluster.
“It’s Toronto AI coming to Boston to absorb biotech capacity,” he said.
The company recently hired three executives who will be based locally and is about to sign a deal on a 10,000-square-foot space in Kendall Square. Between Toronto and Cambridge, Frey said he plans to grow the company from 75 people to more than 160 by the end of 2022.
Deep Genomics is one of many companies using AI for drug discovery, and it expects to have four programs in clinical trials by 2023. The startup also plans to commercialize its own drugs, license its programs to larger companies, and partner with other firms that want to use its AI technology.
Deep Genomics may eventually branch out beyond genetic medicine into diagnostics or digital therapeutics, Frey said. Going public is an option, given the firm’s slate of deep-pocketed investors and Frey’s expectation of where the field is headed in the next 20 years.
“Almost all drugs will be produced using artificial intelligence,” he said. “That’s a change from a focus on chemistry and molecules.” It’s also what leads him to believe Deep Genomics could one day be worth $100 billion.