With a slew of partnerships with large pharmaceutical companies under its belt and the successful spin out of at least one new company, Atomwise has already proved the value of its machine learning platform for discovering and commercializing potential small molecule therapies for a host of conditions.
Now the company has raised $123 million in new funding to accelerate its business.
“Scaling the technology and scaling the team and scaling what we’ve been doing with it,” says chief executive officer Abe Heifets when asked about what comes next for the eight year old business.
Atomwise has already signed contracts worth $5.5 billion with corporate partners that include Eli Lilly & Co., Bayer, Hansoh Pharmaceuticals, and Bridge Biotherapeutics. Smaller, earlier stage companies like StemoniX and SEngine Precision Medicine are also using Atomwise’s tech.
Now the company will look to capture more of the value of drug discovery for itself, looking to develop and commercialize its discoveries by taking over more of the development process and working with manufacturers at a later stage, according to Heifets.
Atomwise tipped its new strategy last year when it announced a partnership with Velocity Drug Development and a $14.5 million investment to create x-37, a spinoff that’s developing small molecule therapies for endodermal cancers, which include cancers of the liver, pancreas, colon, stomach, and bladder.
“We have something like 750 projects running today around the world,” says Heifets. “These comprise more than 600 unique targets and that’s with a vast range of partnerships.”
The power of Atomwise’s drug discovery platform is its ability to harness machine learning to structure new proteins that have never existed — and ensure that they’re able to reach precise target receptors to accomplish the desired task.
Here, the x-37 spinoff is especially illustrative. One line of research the company is conducting into molecules that can target the PIM3 protein receptor. If a drug can block PIM3, it can kill cancerous endodermal cells, according to Heifets. However, if the molecules bind to another, similar target, PIM1, the therapy can cause heart attacks and kill patients.
“This is a challenge and empirically was considered undruggable,” says Heifets. Atomwise’s company screened 11 billion potential molecules against the targets to come up with 500 potential therapies. They’re now working on refining the therapy to bring something to market.
And x-37 is only one of the companies that Atomwise has created to commercialize various new molecules. There’s also Atropos Therapeutics, Theia Biosciences and vAIrus.
Atomwise is far from the only company to think that the application of machine learning technologies to drug discovery is a winning combination. Menten.ai is a company that’s taken the new technology developments one step further and added quantum computing to the mix to come up with new drugs.