Predicting the deleterious impact of variants

The Sorting Intolerant from Tolerant (SIFT) [16, 17] algorithm predicts the effect of a coding variant on protein function, which is displayed as a single measure based on the conservation and scores precomputed and distributed by the J. Craig Venter Institute (http://sift.jcvi.org/). The SIFT Human database that supports GRCh37 Ensembl release 63 (the latest version) was downloaded. In total, 33,248,232 coding DNA sequence variants with a SIFT score, ranging from 0 to 1, and 19,729 protein-coding genes listed in the SIFT Human database were analyzed. Combined Annotation Dependent Depletion v1.3 (CADD; http://cadd.gs.washington.edu/) [18] and Polymorphism Phenotyping v2.2.2 (PolyPhen2; http://genetics.bwh.harvard.edu/pph2/) [19] were also downloaded to predict the deleteriousness of the annotated variants on protein function.