AlphaFold predicts protein structures with atomic accuracy in CASP14 test
Researchers published a deep learning model called AlphaFold in Nature that predicts three dimensional protein structures from amino acid sequences. The neural network draws on physical and biological knowledge about protein structure, including multi-sequence alignments, and its predictions were competitive with experimental structures in a majority of cases. In the 14th Critical Assessment of protein Structure Prediction, CASP14, the system predicted protein structures with atomic accuracy even when no similar known structure existed to guide it. About 100,000 unique protein structures have been determined through decades of experimental work, a small fraction of the billions of known protein sequences, and each structure can take months or years to solve.
Protein structure determines how a protein functions, and computational prediction at this accuracy could help close the gap between the billions of known protein sequences and the far smaller number of experimentally solved structures.
Source: Highly accurate protein structure prediction with AlphaFold - Nature (doi.org).
Written by the Genomes desk from the primary source linked above and checked against it. Research use only; not medical advice. Corrections: [email protected].