DeepMind Releases Third Version Of "AlphaFold" AI Model To Support Drug Discovery

AlphaFold maps DNA, which is designed to help scientists design drugs and target disease more effectively. (photo: x @demissabis)

JAKARTA - On Wednesday, April 8, DeepMind has launched a third version of its "AlphaFold" artificial intelligence model, which is designed to help scientists design drugs and target disease more effectively.

In 2020, the company made significant progress in molecular biology using artificial intelligence that has successfully predicted microscopic protein-protein behavior.

With AlphaFold's latest incarnation, researchers at DeepMind and sister company Isomorphic Labs - both overseen by co-founder Demis Hassabis - have mapped behavior for all molecules of life, including human DNA.

The protein interaction - from enzymes that are essential to human metabolism, to antibodies that fight infectious diseases where other molecules are key in drug discovery and development.

DeepMind said its findings, published in the research journal Nature on Wednesday, would reduce the time and money needed to develop a potentially life-changing treatment.

"With this new capability, we can design molecules that will bind to a specific place on the protein, and we can predict how strong the bonds are," Hassabis said at a press conference on Tuesday, May 7.

"This step is critical if you want to design drugs and compounds that will help treat disease," said Hassabis.

The company also announced the release of "AlphaFold server", a free online tool that scientists can use to test their hypothesis before running real-world trials.

Since 2021, predictions from AlphaFold have been freely accessible to non-commercial researchers, as part of a database containing more than 200 million protein structures, and have been cited thousands of times in other people's works.

DeepMind said the new server required less computational knowledge, allowing researchers to run trials with just a few tap buttons.

"The importance of AlphaFold servers makes it much easier for biologists - who are experts in biology, not computer science - to test larger and more complex cases," said John Jumper, a senior researcher at DeepMind.

Dr. Nicole Wheeler, a microbiologist at Birmingham University, said AlphaFold 3 could significantly accelerate drug discovery line pipes, as it "produces and tests biological designs physically constitutes a large bottleneck in today's biotechnology."