Negative selection of post-translational modification (PTM) regions in human genomes and importance in disease.

<p><b>(A)</b> ∼130,000 experimental PTM sites of four types were merged into ∼55,000 PTM regions. <b>(B-C)</b> Specific negative selection in PTM regions is apparent in relatively higher frequency of rare substitutions and lower ratio of non-synonymous variants to synonymous variants (K_a/K_s). Boxplots represent comparisons of PTM and non-PTM sequence in 100 bins of proteins with matched tolerance to variation. <b>(D)</b> Negative selection of PTM regions is a distinct evolutionary trend not confounded by other genomic factors. PTM-associated predictors are shown in red with the variable corresponding to PTM regions ranked third after conservation and codon bias. <b>(E)</b> Known disease mutations from the HGMD database are enriched in PTM regions. While central PTM sites appear at an expected mutation rate in this global analysis, amino acid weighted sampling reveals an enrichment of PTM sites (<a href="http://www.plosgenetics.org/article/info:doi/10.1371/journal.pgen.1004919#pgen.1004919.s017" target="_blank">S15 Fig.</a>). <b>(F)</b> Disease-associated substitutions in PTM regions are often predicted to be benign by mutation function predictors. Total number of variants scored by each method is shown on each bar.