Estimating the Timing of Multiple Admixture Pulses During Local Ancestry Inference

Abstract

Admixture, the mixing of genetically distinct populations, is increasingly recognized as a fundamental biological process. One major goal of admixture analyses is to estimate the timing of admixture events. Whereas most methods today can only detect the most recent admixture event, here, we present coalescent theory and associated software that can be used to estimate the timing of multiple admixture events in an admixed population. We extensively validate this approach and evaluate the conditions under which it can successfully distinguish one-from two-pulse admixture models. We apply our approach to real and simulated data of Drosophila melanogaster. We find evidence of a single very recent pulse of cosmopolitan ancestry contributing to African populations, as well as evidence for more ancient admixture among genetically differentiated populations in sub-Saharan Africa. These results suggest our method can quantify complex admixture histories involving genetic material introduced by multiple discrete admixture pulses. The new method facilitates the exploration of admixture and its contribution to adaptation, ecological divergence, and speciation.

Using Haplotype Information for Conservation Genomics

“…One main advantage of these methods is that phased data are not needed, and they are also appropriate for low-coverage or pool-sequencing data [84,101]. These methods were tested on simulated data and returned estimates consistent with previous studies on the admixture history of Drosophila melanogaster populations [84,101].…”

“…Novel methods have also been developed to perform local ancestry inference while estimating the timing of a single [84] or several admixture pulses [101], without prior knowledge on the genetic structure of admixture groups [97]. One main advantage of these methods is that phased data are not needed, and they are also appropriate for low-coverage or pool-sequencing data [84,101]. These methods were tested on simulated data and returned estimates consistent with previous studies on the admixture history of Drosophila melanogaster populations [84,101].…”

“…In older hybrid generations, selective effects are more likely to act at the locus scale [87,101]. Maladaptive fitness effects of introgressed alleles could thus emerge only after a long time following hybridization events, when deleterious alleles become dissociated (through recombination events) from each other and from potentially beneficial alleles at other loci.…”

Versatile simulations of admixture and accurate local ancestry inference with mixnmatch and ancestryinfer

“…As local ancestry methods have been applied in more and more species (e.g., Cande, Andolfatto, Prud'homme, Stern, & Gompel, 2012;Li et al, 2019;Sankararaman et al, 2014;Schumer et al, 2014;Slotte et al, 2013), each with their own population genetic properties and demographic histories, simulation tools to evaluate performance have not kept apace. While some studies have carefully modelled the demographic history of hybridizing populations and the impact of this history on the accuracy of ancestry inference (e.g., Medina, Thornlow, Nielsen, & Corbett-Detig, 2018;Sankararaman et al, 2014), this typically requires the development of custom computational pipelines. As a result, many studies do not evaluate the expected performance of local ancestry inference methods and use default parameter sets.…”

The spatial scale of dispersal revealed by admixture tracts

“…Furthermore, we use simulations to evaluate the robustness and the generality of this strategy to different admixture histories and sample sizes. With the development of new LAI methods to estimate the length distribution of admixture tracts (Corbett‐Detig & Nielsen, ; Medina, Thornlow, Nielsen, & Corbett‐Detig, ), we expect that quantitative assessments of dispersal will be obtained in several other species, which may help to resolve a long‐standing issue in conservation biology.…”