Extracting Physiochemical Features for the Prioritization of Candidate Nonsynonymous Single Nucleotide Polymorphisms

摘要

The accelerating advancement of the next generation sequencing technology has been making it feasible to directly sequence candidate genetic regions or the whole genome to obtain rare genetic variants. However, in order to maximize the power of the successive statistical methods for uncovering the possible associations between these rare variants and human inherited diseases, it is desired to include only functional variants in the statistical analysis. To meet this requirement, we propose in this paper to prioritize candidate genetic variants based on the guilt-by-association principle, which assumes that genetic variants associated with the same disease share some common properties. Focusing on a specific type of genetic variants, nonsynonymous single nucleotide polymorphisms (nsSNPs), we take advantages of physiochemical features of amino acids, sequence information of proteins, and multiple sequence alignment of protein families to illustrate the power of prioritizing candidate nsSNPs for specific diseases. Systematic validation experiments show that the proposed approach is capable of effectively recovering the relationship between nsSNPs and diseases, while using the Pearson's correlation coefficient (PCC) to measure the similarity between nsSNPs can achieve the highest performance among all the methods compared.