SAN FRANCISCO & KIEV, Ukraine—In early June, Atomwise Inc. disclosed a multiyear agreement with Eli Lilly and Co. to apply Atomwise’s patented artificial intelligence (AI) technology in support of Lilly’s preclinical drug discovery efforts. The companies will collaborate on up to 10 drug targets selected by Lilly, with the goal of accelerating the time it takes to identify and develop potential new medicines.
According to Abraham Heifets, CEO and co-founder of Atomwise, “There was a strong mutual interest in working together to bring AI-powered drug discovery into the heart of preclinical drug development in Big Pharma. This our first engagement together [with Eli Lilly], but with the excitement we’ve seen thus far in this partnership, it’s unlikely to be the last.”
Lilly is a recognized leader in virtual library design, with a massive number of molecules enabled by automated synthesis in their robotic laboratory. Identifying which molecules might be a potential new therapeutic for specific diseases is a scientific and analytical challenge, one that Atomwise’s AI technology is well suited for.
“We invented the use of a particular kind of machine learning—deep neural networks, which underpin all of the current revolution in practical AI—for structure-based drug design. Our technology uses a statistical approach that extracts insights from millions of experimental binding affinity measurements and thousands of protein structures to predict small molecules-protein binding affinities,” Heifets says. “This fundamental tool makes it possible to do hit discovery and lead optimization and to make toxicity predictions with unparalleled precision. A key feature of the technology is that it makes it possible for our partners to efficiently test a large and diverse set of compounds, which enables the early identification of solutions to potential roadblocks in drug development.”
Heifets also states that “Lilly has made it clear that they are focused on developing drugs for novel target proteins, which are often challenging and less well-studied. Our expertise and tools have been shown to succeed with these kinds of targets, and therefore could be a key to unlocking success for patients.”
Atomwise could receive up to $1 million per target in discovery milestones and will be eligible for up to $550 million in potential development and commercialization milestones inclusive of all targets. As part of the agreement, Atomwise will have the option to develop compounds from the collaboration that Lilly chooses not to advance into clinical testing.
“We want to tackle novel and challenging targets with our current partners, including Eli Lilly, Bayer, UConn and DNDi, and new partners who are working on newly discovered targets,” continues Heifets. “Overall, we aim to make drug discovery more efficient than ever before. We’ll achieve this through co-development efforts with our partners.”
On June 24, Atomwise also announced the launch of a 10-billion compound, AI-powered virtual drug screening initiative: the 10-to-the-10 program, which it is undertaking in collaboration with Enamine Ltd., the world’s largest chemical supplier.
The initiative aims to dramatically increase the discovery of safer small-molecule drugs to treat pediatric cancers. Atomwise will use its AI virtual screening technology to evaluate the binding of billions of drug-like molecules to cancer target proteins, and Enamine will provide support and access to a virtual library of 10 billion small-molecule compounds. The research will be directed by the needs of researchers at leading universities.
“Many of our partners have successfully identified early drug candidates, including submicromolar hits, by screening only 10 million compounds with our AI virtual screening platform,” Heifets tells DDNews. “We’ve barely scratched the surface of what is possible—imagine what will be found when we screen a chemical library that is a thousand times larger.”
The 10-to-the-10 program will look at billions of compounds that have never been examined in any drug discovery program. By evaluating truly novel and structurally distinct compounds, the initiative also dramatically increases the likelihood of developing new drugs for existing targets, with fewer adverse effects.
The enormous screen in the 10-to-the-10 program is possible because of a confluence of technologies: accurate and rapid structure-based drug development with Atomwise’s AI algorithms, scalable cloud computing innovations, and large virtual libraries like Enamine’s REAL (readily accessible) database of compounds that can be synthesized quickly.
Researchers will be able to efficiently test a large and diverse set of compounds, which will enable the early identification of solutions to potential roadblocks in drug development. The initiative aims to not only increase the rate of success, but also to raise the bar for success and advancement at each step of drug development and shorten the time needed for preclinical drug discovery.
Dr. Pengda Liu at the University of North Carolina Lineberger Comprehensive Cancer Center, who has had success with Atomwise’s AIMS Awards, is hoping to take advantage of this program to advance his research.
“In collaboration with Atomwise, we’ve already found a few inhibitors in just one round of screening. Expanding the chemical space to 10 billion could be an absolute game-changer for this research,” said Liu.
The results of this program will be published in peer-reviewed journals.