Inferring the model and onset of natural selection under varying population size from the site frequency spectrum and haplotype structure.

ABSTRACT: A fundamental question about adaptation in a population is the time of onset of the selective pressure acting on beneficial alleles. Inferring this time, in turn, depends on the selection model. We develop a framework of approximate Bayesian computation (ABC) that enables the use of the full site frequency spectrum and haplotype structure to test the goodness-of-fit of selection models and estimate the timing of selection under varying population size scenarios. We show that our method has sufficient power to distinguish natural selection from neutrality even if relatively old selection increased the frequency of a pre-existing allele from 20% to 50% or from 40% to 80%. Our ABC can accurately estimate the time of onset of selection on a new mutation. However, estimates are prone to bias un

SUBMITTER: Nakagome S 

PROVIDER: S-EPMC6408616 | biostudies-literature | 2019 Feb

REPOSITORIES: biostudies-literature