Identifying the Favored Allele in a Selective Sweep

摘要

Abstract Selection is a dominant force in evolution. Mutations arising at random might favor individuals in a specific environmental niche, and populations adapt by rapidly increasing the frequency of individuals carrying the favored mutations. The selection process results in distinct patterns (a signature) of allele frequencies and haplotype structures that can be exploited to identify the genes responding to selection pressure. A study of selection signals in humans has led to molecular insight into the evolution of many natural traits such as skin and eye color, as also adaptation to extreme environments. Computational methods that scan population genomics data to identify signatures of selective sweep have been actively developed, but mostly do not identify the specific mutation favored by the selective sweep. In this talk, we describe an approach that uses population genetics and machine learning techniques to pin-point the favored mutation, even when the signature of selection extends to 5Mbp. Our method, iSAFE, was tested extensively on simulated data and 22 known sweeps in human populations using the 1000 genome project data with some evidence for the favored mutation. iSAFE ranked the candidate mutation among the top 15 (out of ~ 21,000 candidates) in 14 of the 22 loci, and identified previously unreported mutations as favored the 5 regions.