4.1 | Parallel evolution and introgression in caribou ecotypes
Our results confirm previous evidence that northern mountain and boreal caribou from the Northwest Territories are within the BEL genomic lineage, even though they are both currently within the woodland subspecies, confirming that the woodland ecotype appears to have arisen in parallel for both (Polfus et al., 2017). Our central mountain caribou are also within the BEL genomic lineage, and this population has been found to be highly admixed based on the two mtDNA lineages (McDevitt et al., 2009). We found evidence for another, as yet undocumented, case of parallel evolution within the eastern migratory ecotype. The eastern migratory caribou from Ontario/Manitoba and those from Quebec/Labrador are not sister groups (Figures 3b and 4) and have different demographic and introgressive histories.
Recent studies are highlighting that introgression between lineages is far more common than previously realised (Coates, Byrne, & Moritz, 2018; Hamilton & Miller, 2015), and the same appears to be true for caribou with introgression likely playing a role in the evolution of the ecotypes. We find more introgression between caribou from the different lineages than anticipated, for example the barrenground caribou have substantial introgression from the NAL. Introgression is also seen from the NAL into the mountain caribou with an isolation by distance pattern, with negligible levels of introgression into more northerly northern mountain caribou. It thus seems unlikely that introgression drove the parallel evolution of the woodland phenotype of the mountain caribou. High levels of introgression coupled with the finding of many introgressed genes makes a compelling case for parallel evolution through adaptive introgression in the Northwest Territories boreal caribou. However, when we compare the gene compliment of the most highly introgressed regions in the Northwest Territories boreal caribou to those found in the mountain caribou, we again find introgressed regions spread across the genome including many genes, even though there are fewer regions overall. There are a few explanations for this pattern, including incomplete lineage sorting (ILS). ILS would be difficult to exclude, especially given that they are closely related intra-specific ecotypes (Lamichhaney et al., 2017). Whether these regions are a result of ILS or introgression the high gene compliment suggests that they could have persisted in the genome due to selection, even if they have not been involved in the parallel evolution of phenotype, due to filtration for maintenance of adaptive genome segments. Additionally, when studying cases of adaptive introgression in inter-species comparisons, areas of introgression are often restricted to single genomic regions (Schweizer et al., 2018), however in intra-specific taxa we may see larger introgressed regions persisting across the genome because the fitness costs may be lessened. Fully teasing these patterns apart in this case may be complicated because multiple processes are likely acting in concert, including ILS and standing variation being selected upon, coupled with differing levels of introgression as the lineages have come into secondary contact. Additionally, given the PSMC results it is possible that there have been multiple bouts of introgression during glacial cycles over the last ~120,000 years as the lineages repeatedly came into secondary contact. Demonstrating adaptive introgression is complicated and requires the demonstration of the adaptive function of introgressed regions, meaning most cases have thus far have been for well understood traits or those controlled by a single locus (Taylor & Larson, 2019). In contrast, investigating parallel evolution of ecotypes, which will inevitably involve many functional regions, in a non-model species with divergent intra-specific linages and complex demographic histories is a difficult task.