A high-profile Toronto startup looking to use artificial intelligence to discover new drugs for genetic disorders has landed a $13-million (U.S.) private financing led by Silicon Valley venture capitalist firm Kholsa Ventures.
Deep Genomics, led by University of Toronto electrical and computer engineering professor Brendan Frey – one of the top academics in the hot Canadian AI sector – has developed a software system called "Saturn" which it will use to search across 69 billion molecules to identify 1,000 synthesized compounds that can be used to manipulate the makeup of cells.
"Because of the quality of their science and engineering team and the deep integration of their AI technology into their preclinical drug development pipeline, we are confident that a very large potential exists here" to discover new therapies, said Khosla Ventures' billionaire founder Vinod Khosla, a co-founder of Sun Microsystems.
The idea is to target new types of therapies that were previously inaccessible to researchers and apply the resulting molecules in clinical research, with the hope they will eventually be developed into new drugs for genetic ailments. Essentially, the firm is using power of artificial intelligence to speed up and solve the search for the proverbial needle in the haystack.
"We are aiming to flip medicine on its back and do something completely different," said Mr. Frey, 49, who studied at U of T in the 1990s under world-renowned AI "deep learning" pioneer Geoff Hinton, and later worked alongside him.
Mr. Frey came up with the idea of using AI to improve human health in 2002 after researchers successfully sequenced the human genome. Despite that considerable breakthrough, he felt "no human will ever understand the human genome" because it is too complex. With more data and faster computers in the years to come, however, he believed self-teaching algorithms could fill that void by searching huge troves of data to look for common patterns in biology. "I figured the future was developing AI systems to figure out how the genome works."
As the data and computing power caught up, he spent the next 13 years with his research group at U of T working toward his goal, and by 2014 he and his team had published several high-profile scientific papers – including a 2010 cover story in Nature – including the discovery of previously undiscovered biological phenomena. Their AI-based research, for example, showed for the first time how human brain cells spliced toxic proteins to render them inert.
He decided to build a startup to continue his work – rather than continuing to focus on academic research or joining a company – feeling he could have the biggest impact "to change all of medicine. That's what I set out to do."
Deep Genomics generated a lot of buzz when it went through the U of T's Creative Destruction Lab (CDL) startup assistance program in 2015-16, drawing an early $3.7-million (U.S.) investment from Bay-area venture capital firms Bloomberg Beta and True Ventures. True and prominent U.S. biotech investor Bob Nelson are also investing in the company's latest round.
Deep Genomics is "changing the way drug discovery works by applying machine learning techniques to accelerate totally different approach to drug discovery," said CDL head Ajay Agrawal. "They know it will be a long road to revenue, but the prize is so big.
The company has also caught the attention of other high-profile investors, including Canadian astronaut Chris Hadfield and successful Canadian tech entrepreneur and TV personality Michael Hyatt.
"They're aiming to do something that no one else is: building the worlds first and biggest AI-driven genetic medicine company. Not just adding AI to an old fashioned medical company. Creating a company in which AI is everywhere enabling everything."
Mr. Frey said his two-year-old company, which is based in Toronto's MaRS complex, will use the proceeds of the financing to double the size of its 20-person staff, primarily focused on hiring a range of scientific experts.