The Impact of Purifying and Background Selection on the Inference of Population History: Problems and Prospects

Abstract

Current procedures for inferring population history generally assume complete neutrality - that is, they neglect both direct selection and the effects of selection on linked sites. We here examine how the presence of direct purifying selection and background selection may bias demographic inference by evaluating two commonly-used methods (MSMC and fastsimcoal2), specifically studying how the underlying shape of the distribution of fitness effects (DFE) and the fraction of directly selected sites interact with demographic parameter estimation. The results show that, even after masking functional genomic regions, background selection may cause the mis-inference of population growth under models of both constant population size and decline. This effect is amplified as the strength of purifying selection and the density of directly selected sites increases, as indicated by the distortion of the site frequency spectrum and levels of nucleotide diversity at linked neutral sites. We also show how simulated changes in background selection effects caused by population size changes can be predicted analytically. We propose a potential method for correcting for the mis-inference of population growth caused by selection. By treating the DFE as a nuisance parameter and averaging across all potential realizations, we demonstrate that even directly selected sites can be used to infer demographic histories with reasonable accuracy.

Selective sweeps influence diversity over large regions of the mouse genome

“…For example, a population bottleneck may lead to the accumulation of weakly deleterious mutations if drift overwhelms selection. As population size increases after a bottleneck, rapid purging of weakly deleterious mutations can occur, leading to deviations from the expectations of standard models of BGS, which assume constant population size (Torres et al 2020;Johri et al 2021). We have previously inferred a model of demographic history for M. m. castaneus, which suggested that population size has recently increased following a bottleneck (Booker and Keightley 2018).…”

“…We performed an additional set of simulations incorporating this demographic history, but found that the relative reductions in diversity around both protein-coding exons and CNEs were very similar to those observed under constant population size (Figure S4). Note that the trajectory of the demographic history (bottleneck followed by recovery) we inferred may be an artefact of BGS (Ewing and Jensen 2016;Johri et al 2021). However, we proceeded with our analysis assuming estimates of B for a constant population size, because the variations in B around exons and CNEs were very similar with or without population size change.…”

A continent-wide high genetic load in African buffalo revealed by clines in the frequency of deleterious alleles, genetic hitchhiking and linkage disequilibrium

“…According to population genetic theory, high coalescence rates are associated with low heterozygosity because heterozygosity is proportional to the coalescent time [ 73 ]. Although the authors attributed the high coalescence rates in South Africa to a low effective population size, they may also be a signal of selection and hitchhiking effects since the introduction of the aforementioned sex-ratio meiotic gene-drive system [ 74 , 75 ]. If so, then we predict low heterozygosity to be genome-region specific (i.e., near male-deleterious alleles) and not genome wide.…”

Inferring the distribution of fitness effects in patient-sampled and experimental virus populations: two case studies

“…In order to account for these factors, more recent methods have been developed to jointly infer the demographic history simultaneously with the underlying DFE, without making assumptions about the selective effects of any particular class of sites. For example, the approximate Bayesian (ABC) approach recently proposed by Johri et al ( 2020 ) was shown to obtain accurate DFE inference using single time-point datasets, while uniquely accounting for the effects of background selection and the potential non-neutrality of synonymous sites (and see Johri et al 2021 ). Importantly, background selection expectations are themselves incorporated into the inference procedure; as such, the full DFE of newly arising mutations is estimated, even if, for example, the strongly deleterious mutations comprising the most deleterious class are not themselves sampled as polymorphic sites.…”