Selection leads to remarkable variability in the outcomes of hybridization across replicate hybrid zones

Abstract

1AbstractHybrid zones have been held up as an opportunity to see speciation in action, as windows into the evolutionary process. Cases of repeated hybridization provide opportunities for comparative analysis. Within this context, there is an assumption that if the same genetic incompatibilities are maintaining reproductive isolation across all instances of secondary contact, those incompatibilities should be identifiable by consistent patterns across hybrid zones. This is in contrast with changes in allele frequencies due to genetic drift, which will be idiosyncratic for each hybrid zone. To test this assumption, we simulated 20 replicates of each of 24 hybrid zone scenarios with varied genetic incompatibilities, rates of migration, and selection. We found remarkable variability in the outcomes of hybridization in replicate hybrid zones, particularly with Dobzhansky-Muller incompatibilities and strong selection. Specifically, we found substantial differences among replicates in the overall genomic composition of individuals, including admixture proportions, inter-specific ancestry complement in the diploid genome, and number of ancestry junctions. Additionally, we found substantial variation among replicates at focal loci, including a locus that contributed to incompatibility, a locus linked to a locus under selection, and a locus not under selection. From this, we conclude that processes other than selection are responsible for some consistent outcomes of hybridization, whereas selection on incompatibilities can lead to genomically widespread and highly variable outcomes among replicates. We use these results to highlight the challenge of mapping between pattern and process in hybrid zones and call attention to how selection against incompatibilities will commonly lead to unpredictable, variable outcomes. Our hope is that this study informs future research on replicate hybrid zones and encourages further development of statistical techniques, theoretical models, and exploration of additional axes of variation to understand the maintenance of reproductive isolation, while recognizing the difficulties of confidently mapping process to pattern.