The temporal and genomic scale of selection against hybrids
Authors: Jeffrey Groh; Graham Coop
Affiliation: University of California, Davis
Keywords: Speciation & hybridization
Genomic data from hybrid populations contain valuable information about how selection, demography, and gene flow shape divergence and introgression between species. Genome-wide correlations between ancestry state and recombination rate in a growing number of hybrid systems provide evidence of selection broadly acting to remove ancestry of the species contributing the minority of alleles. However, it is unknown to what extent local ancestry patterns reflect selection vs. neutral processes, nor the time-scales of selection on hybrids. To address these questions, we demonstrate the use of the wavelet transform to partition the ancestry variance across loci, as well as the correlation between ancestry and recombination, into contributions from different spatial genomic scales. We use both theory and simulations to illustrate how temporally localized effects of drift and selection are embedded in the spatial decomposition of ancestry variance across loci. We detect signals of strong selection on early generation hybrids in several population genomics datasets. Extending this approach, we estimate the amount of ancestry variance explained by selection across genomic spatial scales, with estimates above 20% at the largest genomic scales. These methods provide a more detailed look into the timescale of selection on hybrids and should be widely applicable as more genomic data sets from hybrid populations are generated.