DeepMind AI Solves 50-Year Protein Folding Challenge

Alphabet's DeepMind team wins a major biennial contest by using its AlphaFold AI to accurately predict the 3D structures of complex proteins, a breakthrough that promises to accelerate disease treatment research.

Alphabet's DeepMind team achieves a massive medical breakthrough as its AlphaFold AI system wins a grand challenge in protein folding. Proteins perform the essential work within our cells, but researchers must understand their complex three-dimensional folds to determine what they actually do. This knowledge is a critical prerequisite for developing many new medical treatments for devastating diseases.

Predicting these 3D structures from a one-dimensional amino acid sequence is a notoriously difficult computational problem that scientists have pursued for fifty years. A typical protein features so many possible structural configurations that brute force calculation would take longer than the age of the universe to solve. While traditional experimental methods like X-ray crystallography provide answers, they remain slow and expensive processes that delay vital medical research.

AlphaFold successfully tackles this challenge by predicting protein structures with near-atomic accuracy in two-thirds of the test cases during the biennial CASP contest. By providing a fast and highly accurate computational alternative to traditional methods, this AI system gives researchers a powerful new tool. This capability ultimately allows the medical community to analyze cells and viruses much more quickly as they work to understand and combat various diseases.

Read More at the original source →