Machine-learning technology has beaten people at games of chess and Go to worldwide fanfare. An indication of its eerily lifelike prowess in making phone calls to unsuspecting folks went viral.
However, a less-noticed win for DeepMind, the artificial intelligence arm of Google’s parent Alphabet Inc., at a biennial biology conference might upend how drugmakers find and develop new medicines. It might additionally dial-up pressure on the world’s largest pharmaceutical corporations to organize for a technological arms race. Already, a new breed of upstarts is jumping into the fray.
In December, on the CASP13 assembly in Riviera Maya, Mexico, DeepMind beat seasoned biologists at predicting the shapes of proteins, the essential building blocks of disease. The seemingly esoteric pursuit has severe implications: A tool that may precisely model protein structures could speed up the development of new drugs.
“Stunning,” tweeted one scientist after the raw outcomes were posted online. “It was a complete surprise,” stated conference founder John Moult, a University of Maryland computational biologist. “In comparison with the history of what we had been able to do, it was fairly spectacular.”
Sorting out the construction of proteins with a purpose to discover ways for medicines to attack disease is an enormously complex problem. Researchers still don’t fully perceive the rules for the way proteins are constructed. After which there’s the mathematics: There are more possible protein shapes than there are atoms within the universe, making prediction a problematic undertaking of computation. For a quarter-century, computational biologists have labored to devise software equal to the task.
Enter DeepMind. With limited experience in protein folding — the physical process by which a protein acquires its three-dimensional shape — however, armed with the latest neural-network algorithms, DeepMind did more than what 50 top labs from around the world could accomplish.