Google spent much of last year scrambling to build its Gemini chatbot to counter ChatGPT, pitching it as a multifunctional AI assistant that can help with work tasks or digital tasks in personal life. More quietly, the company has been working to improve a more specialized artificial intelligence tool that is already a must-have for some scientists.
AlphaFold, software developed by Google’s DeepMind AI unit to predict the 3D structure of proteins, has received a significant update. It can now model other biologically important molecules, including DNA, and the interactions between antibodies produced by the immune system and molecules from pathogenic organisms. DeepMind added those new capabilities to AlphaFold 3 in part by borrowing techniques from AI imagers.
“This is a big breakthrough for us,” Demis Hassabis, CEO of Google DeepMind, told WIRED ahead of Wednesday’s release of a document on AlphaFold 3 in the scientific journal Nature. “This is exactly what is needed for drug discovery: you have to see how a small molecule will bind to a drug, how strongly and also what else it might bind to.”
AlphaFold 3 can model large molecules such as DNA and RNA, which carry the genetic code, but also much smaller entities, including metal ions. It can predict with great accuracy how these different molecules will interact with each other, the Google research paper states.
The software was developed by Google DeepMind and Isomorphic Labs, a sister company of parent company Alphabet that works on artificial intelligence for biotechnology and is also led by Hassabis. In January, Isomorphic Labs announced it would work with Eli Lilly and Novartis on drug development.
AlphaFold 3 will be available through the cloud for outside researchers to access for free, but DeepMind will not release the software as open source as it did with previous versions of AlphaFold. John Jumper, who leads the Google DeepMind team working on the software, says it could help provide a deeper understanding of how proteins interact and function with DNA within the body. “How do proteins respond to DNA damage? How do they find it and repair it? Jumper says. “We can begin to answer these questions.”
Understanding protein structures used to require painstaking work using electron microscopes and a technique called predict the shape of proteins simply from their constituent amino acids, learning from structures that had been verified experimentally.
In 2018, Google DeepMind revealed that it was working on artificial intelligence software called AlphaFold to accurately predict the shape of proteins. In 2020, AlphaFold 2 produced sufficiently accurate results to unleash a storm of enthusiasm in molecular biology. A year later, the company released an open source version of AlphaFold for anyone to use, along with 350,000 predicted protein structures, including almost every protein known to exist in the human body. In 2022, the company released more than 2 million protein structures.
