As more and more biopharmas develop artificial intelligence platforms, the drug discovery process is being reshaped to fit new goals on cutting down the prodigious amount of time, energy and money that go into a drug program. Now one of the most ambitious players in the drive to improve on ROI, AstraZeneca, is marking a milestone on that front by adding the first target generated by AI to its portfolio.
The target comes out of a collaboration with BenevolentAI, a London-based company that first started working with the big pharma back in April, 2019. Thanks to the biotech’s platform AI tech, AstraZeneca has identified and validated a new way to attack chronic kidney disease and will begin developing compounds centered on this target.
This selection fits into a broader AI strategy at AstraZeneca, the company’s renal biosciences chief told Endpoints News. And the plan isn’t to just limit its applications to R&D.
“We’re investing in AI and see it as a clear tool that can support our decisions in drug discovery,” Pernille Hansen said. “They can be not only for discovering new targets, it can be in chemistry, imaging and so forth. So there’s many, many possibilities to work with AI.”
But they’re still a long way out from proving AI can offer tangible evidence of an impact.
Hansen declined to say what specifically the target is, what kinds of drugs will utilize it to treat CKD or the timeline for such drug development. The only thing she’d disclose is that it suits their renal and AI approaches, and that AstraZeneca researchers have given the OK to advance.
Kidney disease is a familiar area for the company, given its studies to expand its blockbuster Farxiga drug into CKD over the last several years. A Phase III trial in the disease was halted in March over what AstraZeneca deemed “overwhelming efficacy,” and the company released topline results in July. The FDA has set a rough review date for sometime during the second quarter.
If Farxiga does cross the finish line in CKD, it would represent a second approval once thought unlikely for a drug class developed for diabetes. The drug was approved in May to reduce the risk of cardiovascular death or hospitalization in heart failure patients with a reduced ejection fraction, and those with or without type 2 diabetes.
Wednesday’s target, though, comes out of BenevolentAI’s platform, which essentially forms the shape of a “knowledge graph,” COO Ivan Griffin told Endpoints News. Griffin’s company has spent years feeding its tech with all sorts of data, from proteins and genes to results published in scientific journals, and training its algorithm to make connections that scientists may not have noticed at first.
One could visualize such the graph as an interconnected web of “nodes and edges,” Griffin said. Scientists take advantage of the AI-predicted relationships and then interrogate them to see if it holds up, which is what happened in AstraZeneca’s case.
“The strengths we have in the data and AI and machine learning technology approach matches really nicely, in this case, with what AstraZeneca was able to bring,” Griffin said. “By combining the two, and then setting off together to try to discover targets and drugs, that was the philosophy behind starting the collaboration, and it’s very much felt like that all the way through.
In addition to CKD, BenevolentAI and AstraZeneca are working on developing drugs for idiopathic pulmonary fibrosis as part of their collaboration.