Selective pressure at the codon level improves the prediction of disease related protein mutations in human

Motivation: Predicting the functional impact of a protein variation is one of the most challenging problems in Bioinformatics. A rapidly growing number of genome-scale studies provide large amounts of experimental data allowing the application of rigorous statistical approaches for predicting if a given single point mutation has or not an impact on human health. Up until now, existing methods have limited their source data to either protein or gene information. Novel in this work, we take advantage of both and focus on protein evolutionary information by using estimated selective pressures at the codon level. Results: Here we introduce a new method (SeqProfCod) to predict the likeliness that a given protein variant is associated or not with a human disease. We have developed a Support Vector Machine classifier trained using three sources of information: protein sequence, multiple protein sequence alignments and the estimation of selective pressure at the codon level. SeqProfCod has been benchmarked with a large dataset of 9,979 single point mutations from 1,599 human proteins from SWISS-PROT. It achieves 77% overall accuracy and a correlation coefficient of 0.49 indicating that the estimation of the selective pressure helps in predicting the functional impact of single-point mutations. Thus, this study demonstrates the synergic effect of combining the two classical sources of information for predicting the functional effects of protein variants: protein sequence/profile-based information and selective pressures at codon level.