Proteins are immensely important for the operation of cells, and predicting how they go from sequences of their amino acids to compact shapes that drive life is central to biology and medicine.
This year’s prestigious Breakthrough Prize in life sciences has been awarded to two scientists, Demis Hassabis and John Jumper for developing a computational tool using artificial intelligence that has largely solved the protein structure problem. Hassabis and Jumper are leaders behind AlphaFold 2, the system they developed with their team of scientists and engineers at DeepMind in London.
AlphaFold is a deep-learning program that accurately and rapidly models the structure of proteins. It already has a revolutionary impact in the life sciences: This summer, DeepMind has uploaded the structures of 200 million proteins unraveled by their program – nearly every known protein from across the tree of life – to a public database. The database will reduce the time scientists spend determining protein structure from months or years to hours or minutes, opening the door for many exciting future discoveries, from drug design to synthetic biology, the development of nanomaterials, and the fundamental understanding of life’s cellular processes.
The inside story on how the team at DeepMind created AlphaFold to solve the problem of protein folding.
This post is adapted from the Breakthrough Prize website announcing the winners of this year’s prizes in life sciences, mathematics, and fundamental physics. Visit https://breakthroughprize.org/ for more stories.