The Temporal Dynamics of Background Selection in Nonequilibrium Populations

Raul Torres,Ryan D Hernandez,Markus G Stetter,Jeffrey Ross-Ibarra

doi:10.1534/genetics.119.302892

Abstract

Neutral genetic diversity across the genome is determined by the complex interplay of mutation, demographic history, and natural selection. While the direct action of natural selection is limited to functional loci across the genome, its impact can have effects on nearby neutral loci due to genetic linkage. These effects of selection at linked sites, referred to as genetic hitchhiking and background selection (BGS), are pervasive across natural populations. However, only recently has there been a focus on the joint consequences of demography and selection at linked sites, and some empirical studies have come to apparently contradictory conclusions as to their combined effects. To understand the relationship between demography and selection at linked sites, we conducted an extensive forward simulation study of BGS under a range of demographic models. We found that the relative levels of diversity in BGS and neutral regions vary over time and that the initial dynamics after a population size change are often in the opposite direction of the long-term expected trajectory. Our detailed observations of the temporal dynamics of neutral diversity in the context of selection at linked sites in nonequilibrium populations provide new intuition about why patterns of diversity under BGS vary through time in natural populations and help reconcile previously contradictory observations. Most notably, our results highlight that classical models of BGS are poorly suited for predicting diversity in nonequilibrium populations.

Highlights

The effects of natural selection and demography on neutral genetic diversity within populations have long been of interest in evolutionary and population genetics
Selection parameters for simulating background selection (BGS) followed those of Torres et al (2018), with deleterious variation occurring at 20% of sites across a 2 Mb locus and the selection coefficient, s, drawn from two distributions of fitness effects (DFEs)
What considerations should researchers interested in empirical analysis keep in mind? While simulation results suggest that BGS is unlikely to strongly affect the ability to detect outliers via selection scans using FST (Matthey-Doret and Whitlock 2019), we argue here against using simple approximations based on equilibrium models to infer the relative importance of demography and selection in patterning diversity along the genome

Summary

Introduction

The effects of natural selection and demography on neutral genetic diversity within populations have long been of interest in evolutionary and population genetics. While the theoretical underpinnings of the effects of natural selection and demography on genetic diversity have been investigated for decades (Maynard Smith and Haigh 1974; Nei et al 1975; Maruyama and Fuerst 1984, 1985; Kaplan et al 1989; Tajima 1989; Charlesworth et al 1993; Hudson and Kaplan 1995; Nordborg et al 1996), detailed investigation into how they jointly act to create patterns of diversity in different populations remains lacking Both theory and empirical observation have long shown that patterns of neutral genetic variation can vary regionally across the genome as a function of recombination rate (Maynard Smith and Haigh 1974; Begun and Aquadro 1992). Evidence for selection at linked sites has been found across an array of species, including Drosophila melanogaster (Begun and Aquadro 1992; Charlesworth 1996; Andolfatto 2007; Sella et al 2009; Comeron 2014; Elyashiv et al 2016), mice (Keightley and Booker 2018), wild and domesticated rice (Flowers et al 2011; Xu et al 2012), Capsella (Williamson et al 2014), monkeyflowers (Stankowski et al 2018), maize (Beissinger et al 2016), and humans (Sabeti et al 2002; Reed et al 2005; Voight et al 2006; McVicker et al 2009; Cai et al 2009; Hernandez et al 2011; Lohmueller et al 2011)

Methods

Results

Discussion

Conclusion