(Photo by University of Oxford/Getty Images)
Predicting how proteins fold up has been a problem that scientists have been trying to solve for generations, but DeepMind, an artificial intelligence research lab that's a part of Google's parent company Alphabet, developed an AI system that can solve these problems in hours instead of years, according to Vox.
The upshot could be faster development of radically more effective drugs and treatments for illnesses such as Alzheimer's disease as well as the ability to more quickly determines which existing drugs will be most effective against particular illnesses, which could be essential in a rapidly changing situation like a pandemic.
The Critical Assessment of Structure Prediction (CASP) conference announced Monday that AlphaFold, the DeepMind AI system, was able to predict the 3D structures that proteins fold into to within one atom of accuracy.
That's significant because the folded-up structure determines how a protein works. If scientists know how to predict the folds and therefore the structures, they can better develop molecules that can attach to the protein to alter its function, leading to novel treatments.
John Moult, who co-founded CASP, told Nature, "This is a big deal," adding, "In some sense the problem is solved."
Andrei Lupas, an evolutionary biologist who was involved with this year's CASP conference, used AlphaFold to find the structure of a protein that his lab has been trying to find for a decade.
"It's a game changer," he said. "This will change medicine. It will change research. It will change bioengineering. It will change everything."