Deep Genomics, UTEST cohort 2015, a Toronto-based AI therapeutics startup, has made the first-ever discovery of a diseases treatment and drug candidate using artificial intelligence.
The startup announced on Wednesday that its propriety AI-based drug discovery platform has identified a novel treatment target and corresponding drug candidate for Wilson disease, a rare and potentially life-threatening genetic disorder.
“Our AI systems can figure out how diseases are caused and how to fix those diseases much more rapidly than humans ever could.”
“This is the amazing accomplishment for the team,” Brendan Frey, founder and CEO of Deep Genomics, told BetaKit. He noted that part of Deep Genomic’s goal is to “help everyone in the world” use the discovery and technology its developing in Canada to support discovery and development more broadly.
A professor of electrical and computer engineering at the University of Toronto, Frey said he first began building out the “primitive” versions of Deep Genomics’ AI model in 2004. Over the next 10 years, he and his team built more and more advanced AI systems and in 2014 built a system that could interpret human genetic mutations more accurately than humans could.
It was at that point that Frey figured the corporate medical world would become interested in the work and “run with it.” But that didn’t happen, so Frey decided to do it himself, founding Deep Genomics in 2015.
Frey called Deep Genomics unique in that its artificial intelligence model accurately understands how biology works. “Our AI systems can figure out how diseases are caused and how to fix those diseases much more rapidly than humans ever could and also with a greater success rate,” Frey explained.
A new era of drug discovery
Deep Genomics’ platform uses deep learning-based technology to help geneticists, molecular biologists, and chemists develop therapies. Founded in 2015, the company developed a software system called Saturn, which it uses to search across 69 billion molecules to identify 1,000 synthesized compounds that can be used to manipulate the makeup of cells.
“I am certain that we are witnessing a new era of drug discovery.”
The treatment and drug candidate for Wilson marks the first successful drug discovery on the platform, and the first AI-discovered drug candidate globally, with Deep Genomics beating out other companies, including pharmaceutical giants, working towards the same goal.
“This is truly unprecedented and opens the door to a smarter, faster, and vastly more efficient means of identifying viable drug candidates for a host of diseases,” said Arthur A. Levin, a member of Deep Genomics strategic advisory board. “Developing new therapeutics is full of unknowns, but I am certain that we are witnessing a new era of drug discovery.”
Deep Genomics will develop the candidate, called DG12P1, for the treatment of patients with Wilson disease, which causes a genetic variant that impairs the body’s ability to remove copper. Wilson disease affects approximately one in every 30,000 people worldwide and, if left untreated, can cause organ damage.
“Researchers have struggled for two decades, without success, to understand the mechanism of this genetic variant that causes Wilson disease,” said Dr. Frederick K. Askari, an associate professor and director of the Wilson disease program at the University of Michigan. “The clarity that this artificial intelligence platform has brought to the scientific community is astounding and the potential of a therapy that could operate at the genomic level to correct the disease process is exciting.”
Breaking it down from years to months
“Drug discovery and development is normally a very lengthy process,” explained Frey. “Going from a patient with the disease and trying to figure out how the disease works all the way through to getting an approved drug takes 17 years.”
By using AI, the process to find the drug compound took Deep Genomics less than 18 months.
“Nowadays, there’s also these increasing amounts of data available, different sensor measurements, different kinds of experimental data. There’s been a mass explosion of data, which is a double-edged sword,” Frey told BetaKit. “From one perspective, it can help us find drugs more quickly and with the greater success rates, but on the other hand, it’s just a lot of information and humans aren’t good at sorting through the information.”
“By finding these drugs more rapidly, and advancing more quickly, [Deep Genomics] can save a lot of money.”
Frey stated that Deep genomics breakthrough is in developing AI systems that can sort through all that excessive data and “very clearly” figure out how the biology works and how to design a compound. He told BetaKit that the company is currently working to discover other treatments and drug candidates and plans to declare another five to seven candidates in the next year and a half.
The next steps for DG12P1 is to obtain an investigational drug in the first quarter of 2021, and then pursue regulatory approval globally so the treatment can be used by humans.
Frey also noted that Deep Genomics, which has raised $21 million to date, is looking to raise additional funding in the fall of 2019 in order to grow the company. The startup will look to grow its team, which currently sits at 40 people, as well as support the expensive process of drug discovery.
However, Frey explained that Deep Genomic’s method is actually more cost effective than traditional approaches.
“One of the biggest costs in drug development is time,” he said. “When it takes a long time, it takes a lot of money. By finding these drugs more rapidly, and advancing more quickly, [Deep Genomics] can save a lot of money.”