Deep Genomics, now flush with cash, plans to take dozens of RNA therapies to the clinic
It was 2002 when Brendan Frey noticed a huge gap in biotech. The human genome had just been sequenced, allowing scientists to map genetic mutations. But there weren’t enough data to understand the consequences of those mutations, or really do much about them.
Predicting there would be an explosion of new data, Frey spent the next 13 years working on a way to sift through it all. Now, thanks to advances in RNA therapeutics, medicine is becoming programmable, Frey said. And on Wednesday, a slate of investors bet $180 million that his company’s AI platform can make sense of it.
“What’s really cool about RNA therapeutics is that they’re basically a sequence of letters,” Frey said. “Change the sequence of letters one way, you target a different gene. Change it one way, you can increase the amount of protein produced by that gene. Change the sequence of letters a different way, you can decrease the protein.”
It’s all digital information, Frey said. And thanks to AI and deep learning tools, Deep Genomics says it can do things like figure out which mechanisms of action will (or won’t) work against a specific gene, without performing a single experiment.
“We can take a gene where another company would have spent two years on it and then failed and dropped it, and we can actually drug that gene, or we know to put it at the bottom of the list, just don’t do it now, it’s going to be too hard,” he said.
In 2019, the company put forward its first preclinical candidate, a therapy for Wilson disease that’s expected to hit the clinic along with three other candidates by 2023. Using the AI system, the team says it was able to go from target identification to declaring a winner in 18 months. Deep Genomics has a total of 10 candidates hurtling toward the clinic, and Frey says he expects to add 20 more in the near future.
The other three candidates expected to hit the clinic by 2023 are for frontotemporal dementia, gout and Niemann-Pick type C disease.
The AI space is teeming with players, like Enveda, which nabbed a $51 million Series A round last month to pursue new therapies for Wilson disease, NASH and Parkinson’s disease. Upon pulling in a $225 million Series C round last month, Insilico CEO Alex Zhavoronkov laid out big plans to emerge as the Amazon or Google of the field. Around the same time, UK-based Exscientia splurged on the three-year-old molecule screening biotech Allcyte in an attempt to edge out rivals.
What separates Deep Genomics from some of its peers — like Recursion, Exscientia or insitro — is its sole focus on RNA biology, Frey said.
“We like RNA biology because of that rock solid framework,” he said. “We have 100 petabytes of data, so everything’s in place.”
Frey says the AI tech is less like a grand, all-knowing computer, and more like a workbench, with dozens of tools that have defined functions and scopes. For example, one tool was built to go through databases of patient mutations and find drug targets based on RNA biology — but researchers at the company realized it could also be used to analyze different types of animal models and figure out which ones would recapitulate human biology.
“That’s kind of the advantage of the workbench metaphor is it sort of frees people up to be more creative,” Frey said.
Softbank Vision Fund 2 led the Series C round, with a hand from Fidelity Management & Research Company, Canadian Pension Plan Investment Board, True Ventures, Amplitude Ventures, Khosla Ventures and Magnetic Ventures. When asked if an IPO is in the near future, Frey said he doesn’t plan to take the company public at least until they reach the clinic.
“A lot of companies have gone public preclinically in the last year, and the problem is that if they stumble in getting into the clinic, then the existence of the company will be put into question,” he said. “We don’t want to be in that situation.”
