Predicting the phenotype of Mendelian disease missense mutations using amino acid conservation and protein stability change

Abstract

Many Mendelian diseases are caused by recessive, loss-of-function missense mutations. On a gene-by-gene basis, it has been demonstrated that missense mutations cause, among other defects, protein misfolding, protein instability, protein mistransport, which strongly suggests that pathogenic missense mutations do not occur at random positions. Based on those observations, we predicted that Mendelian disease missense mutations are enriched in evolutionarily-conserved amino acids. In a pilot set of 260 Mendelian diseases genes affecting cellular organelles we show that missense mutations indeed occur in amino acids that are significantly more conserved than the average amino acid in the protein based on three different scoring methods (Jensen Shannon Divergence p = 7.78E-03, Shannon Entropy p = 1.68E-13, Sum of Pairs p = 1.55E-17). In order to understand how these results might be related to clinical phenotypes in humans or preclinical phenotypes in model organisms, we calculated the protein stability change upon mutation ({Delta}{Delta}Gu) using EASE-MM and found that, on average, pathogenic mutations cause a stability change of greater magnitude than benign mutations (p = 4.414428E-23). Finally, we performed a computational case study on NPC1, the gene responsible for 95% of diagnosed cases of the lysosomal storage disorder Niemann-Pick Type C using a set of 411 missense mutations from the Exome Aggregation Consortium.