Protein Structure Prediction to Function Prediction Using AlphaFold

Abstract

AlphaFold which builds in deep learning-based methods based on large amounts of protein data allows us highly accurate structure prediction/modeling for most proteins. AlphaFold models of proteins can be useful for diverse applications such as predictions of intrinsically disordered regions, modeling of structures based on cryo-EM maps and ligand binding site predictions. Here, we introduce and review the backgrounds, effects and caveats of protein structure/complex structure prediction using AlphaFold/AlphaFold-Multimer.