For years, experts have pointed out that AI could make it faster and cheaper to find new drugs to treat a variety of diseases. For example, AI could help scan databases of potential molecules to find some that are best suited to specific biological targets, or fine-tune proposed compounds. Companies developing ai tools have received hundreds of millions of dollars in investment over the past two years.
A spokesperson said in a statement to The Verge that The company will focus on developing partnerships with pharmaceutical companies.
However, developing and testing drugs can be a bigger challenge than figuring out the structure of proteins. For example, even when the structures of two proteins fit physically, it is hard to tell whether they will stick together at all. A drug candidate based on how it works at the chemical level may not always be effective when given to animals or people.
As chemist and author Derek Lowe pointed out in Science this summer, more than 90 percent of drugs that enter clinical trials end up not working. Most of the problems are not at the molecular level.
DeepMind's work and Isomorphic's planned work may help break through some research bottlenecks but are no quick solutions to the myriad challenges of drug development. Helen Walden, a professor of structural biology at The University of Glasgow, previously told The Verge that "The laborious, resource-consuming work of doing biochemical and biological assessments of things like drug function" will continue.