Genomic analysis of AlphaFold2-predicted structures identifies maps of 3D essential sites in 243 neurodevelopmental disorder-associated proteins

Neurodevelopmental disorders (NDDs) are heterogeneous, congenital conditions, often caused by missense variants with varying, hard-to-elucidate molecular effects (e.g., gain- or loss-of-function). Functionally indispensable regions in three-dimensional (3D) structures of proteins encoded by associated genes can provide insights into the molecular mechanisms of disease variants. The recent success of AlphaFold2 in predicting protein structures at an accuracy matching experimental methods opened an avenue to computationally identify 3D essential sites for disease genes with no available experimental structure (67%).