Home Health Care Relay Therapeutics pays $85M for startup with a new AI tech for...

Relay Therapeutics pays $85M for startup with a new AI tech for drug discovery

4
0
SHARE

 

Relay Therapeutics has two cancer drugs in early clinical development and one more close behind—all of them discovered and developed with proprietary computational technology that finds ways to drug hard-to-hit protein targets. That technology is producing more drug candidates, but Relay is expanding its drug discovery reach by buying an artificial intelligence startup that brings a new dimension to its search for new molecules.

Cambridge, Massachusetts-based Relay is acquiring privately held ZebiAI for $85 million. According to deal terms announced Friday, Relay is paying $20 million cash up front and $65 million in the publicly traded company’s stock.

Relay’s own technology, called Dynamo, uses artificial intelligence and machine-learning techniques to analyze protein motion, which is the way that proteins change shape and the role a particular shape plays in disease. That analysis is then used to find novel sites on a protein for small molecule to bind to, places that have haven’t been hit before.

ZebiAI is also working to identify small molecules that bind to a protein. The startup works with DNA-encoded libraries (DELs), collections of small molecules that are each tagged with a unique DNA sequence, said Relay CEO Sanjiv Patel, speaking on an investor call. Unlike DEL models that focus only on the hits, ZebiAI factors in molecules that don’t bind to the proteins. All of the data are used to train machine-learning models, which have been built in collaboration with the Accelerated Sciences Group of Google.

Patel said that ZebiAI’s approach can predict potential small molecules that bind to a protein of interest. The technology has the potential to more rapidly find those molecules, and it creates a process that doesn’t require the physical making of a molecule until the research is close to drug-like starting point. Patel said that this machine learning-DEL (ML-DEL) approach makes drug discovery more efficient, robust, and effective.

“The more ML-DEL screens that are done, and the more predictions made, and the more they are validated, the better all this gets,” Patel said. “This is a perfect complement to Relay Therapeutics’ Dynamo platform.”

ZebiAI has used its technology to validate several targets. The company published results of its research last June in the Journal of Medicinal Chemistry.

With ZebAI’s acquisition, Rafael Gomez-Bombarelli, chief learning officer of ZebiAI and a professor at Massachusetts Institute of Technology, will join Relay as an advisor. The data that ZebiAI generated are now owned by Relay, as of the close of the acquisition. If ZebiAI’s research proves successful, ZebiAI’s shareholders could earn up to $85 million in milestone payments, payable in Relay stock, according to the deal terms. If Relay starts working with other companies in partnerships based on ZebiAI’s technology, ZebiAI shareholders could earn 10% of the payments that Relay receives from those collaborations within the next three years. Those payments could reach up to $100 million cash.

Meanwhile, Relay is making progress with compounds discovered with Dynamo. Lead program RLY-1971 is a small molecule that selectively targets SHP2, a cancer protein in the RAS pathway that promotes the survival and growth of cancer cells. The Relay drug is designed to bind to and stabilize SHP2 in its inactive shape. The company is currently testing the drug in a Phase 1 study. RLY-1971 is being developed under a partnership with Genentech that started late last year. Patel said that a clinical trial testing a combination of RLY-1971 with Genentech’s KRAS inhibitor, GDC-6036, is on track to start later this year.

The next program in Relay’s pipeline, RLY-4008, is a small molecule that targets FGFR2, a protein that is often mutated in cancers. Though there are cancer drugs available that block FGFR2, they also affect other proteins. Relay designed its FGFR2 inhibitor to selectively block the target protein while minimizing its effect on other proteins. The drug is currently in a dose-escalation Phase 1 study in patients whose solid tumors have FGFR2 alterations. Preclinical data were presented at the recent American Association for Cancer Research.

Photo: Andrzej Wojcicki, Getty Images

Source link

LEAVE A REPLY

Please enter your comment!
Please enter your name here