Proteins are the basic building block of life. Each protein is composed of a sequence of amino acids. The function of a protein is largely determined by its structure; however, determining structure from sequence is a challenging problem which computational biologists have attempted to solve for over five decades (see the Critical Assessment of Protein Structure Prediction [CASP] experiments).
In November 2020, a team from Google DeepMind released AlphaFold 2 (AF2), an artificial intelligence (AI) system to predict 3D protein structure from an amino acid sequence. AF2 won the CASP14 competition, making the best prediction for 88 out of the 97 targets and significantly outperforming other competitor algorithms. AF2 achieved a median global distance test (GDT) score of 92.4/100 and a median root-mean-square deviation (RMSD) of 2.1 Å. These accuracy scores are competitive with often prohibitively time and resource-expensive experimental approaches such as X-ray crystallography. Further, AF2 demonstrated impressive performance at modeling side chain orientations.
"In this study, we develop the first, to our knowledge, computational approach capable of predicting protein structures to near experimental accuracy in a majority of cases."
AF2 is poised to revolutionize both structural biology and its applications across myriad domains of biology. This website tracks publications pertaining to AF2 and its applications. Further, the Examples page offers several interactive comparisons between experimentally-determined structures and AlphaFold 2 predictions.