CASP (Critical Assessment of Structural Prediction) is a competition for computational biologists to develop methods to blindly predict the structure of proteins as close to their experimental data. The reason for the jubilation and subsequent words of caution from the structural biology community is that Google’s AlphaFold2 managed to accurately predict the structure of the CASP14 protein from this year’s competition, head and shoulders above other modeling methods. On the CASP Global Distance Test (GDT) - a measure of amino acids in the protein within a threshold distance from the correct position, AlphaFold2 scored an impressive 90, competitive with experimental results.
Why is protein structure important?