Speciation results from the accumulation of genetic differences between lineages over time, which initially decreases and eventually eliminates the probability of gene flow between them (i.e., biological species concept). Additionally, the genomes of natural populations are not only shaped by drift and selection, but also by introgression from closely related taxa following secondary contact. Secondary contact between distinct, yet interfertile, lineages may lead to outcomes ranging from complete unification to formation of stable, narrow hybrid zones permitting low levels of genetic exchange. These stable hybrid zones can be maintained either by pre-zygotic (e.g., behavioral) or post-zygotic (e.g., reduced hybrid fitness) reproductive barriers. Additionally, gene flow following long-term isolation provides opportunities for selective introgression between lineages. The process of speciation is therefore a continuum, and there is regular debate as to the classification status of related lineages that have not yet reached complete reproductive isolation and are instead in the “gray zone” of speciation.
Secondary contact between lineages can result from either natural or anthropogenic forces. For example, stable hybrid zones typically arise from the natural expansions and contractions of lineages throughout geologic time due to major climatic fluctuations. Over the past century, however, human translocations have become prevalent, sometimes leading to secondary contact between introduced and native populations of the same species. Here I investigate the dynamics of secondary contact and hybridization between distinct canid lineages, focusing on two different systems, the red fox (Vulpes vulpes) and the gray fox (Urocyon cinereoargenteus). These vary both in the origins of secondary contact (anthropogenic vs. natural range expansion) and in the level of divergence between lineages (late-Pleistocene vs. mid-Pleistocene), making them valuable systems to explore the mechanisms maintaining lineage boundaries and the role of selective introgression in their evolutionIn Chapter 1, I investigated patterns of human facilitated gene flow between two lineages that are >20,000 years divergent. The native Sacramento Valley red fox (SVRF, V. v. patwin) is endemic to the semi-arid region of California’s northern Central Valley. In direct contact with the SVRF range is a population of nonnative red foxes, found primarily in the San Joaquin valley to the south of the native population and in the coastal lowland region to the west. This nonnative population was derived from multiple human translocations of fur-farmed foxes in the early 1900s. Most farmed foxes were originally sourced from eastern Canadian and Alaskan lineages in the late 1800s which were phylogenetically divergent (~20–70 kya) from the SVRF. They were bred in fur farms for several decades prior to their release or escape in California in the mid 1900s. I hypothesized that gene flow was restricted, potentially due to post-zygotic genetic mechanisms, and that some genes originating in nonnative foxes would confer higher fitness in the currently human-dominated landscape and would therefore be selectively introgressed into the native fox population. I sequenced 107 red foxes from the native (n = 59) and nonnative (n = 48) ranges at a mitochondrial fragment and >19,000 loci of the nuclear genome.
Observed geographic cline widths were 6.9× (mtDNA) and 14.3× (nuDNA) narrower than expected based on simulations assuming unrestricted gene flow, consistent with the presence of reproductive barriers. Using a Bayesian genomic cline analysis, I identified 10 loci with significantly reduced levels of introgression, several of which were previously associated with reproductive fitness. Consistent with selective introgression, nine loci were identified with significantly elevated levels of gene flow, most of which originated from the nonnative population. Several genes near these outlier regions were potentially associated with adaptation to human dominated landscapes. It should be noted that pre-zygotic factors, such as assortative mating or natal habitat-biased dispersal, also could have contributed to the maintenance of the hybrid zone. Nevertheless, these findings indicate the presence of some form of reproductive barrier between the native and nonnative red fox populations, which enabled the identification of several exceptional genes that were shared at much higher rates than expected by chance. These genes flowed primarily from the nonnative population, for which ancestors had undergone strong selection for a captive environment, to the native population, which only recently (150 years) experienced the conversion of its historical range to a human-dominated landscape.
In Chapters 2 and 3 I investigated patterns of divergence and gene flow between two lineages that are ~1 million years divergent, where secondary contact was presumed to be a result of natural range expansion. North American gray foxes are composed of two highly divergent, reciprocally monophyletic lineages in the western and eastern portions of their range. They currently hybridize in a relatively a narrow zone of contact in the southern Great Plains. The narrowness of their hybrid zone indicates either that secondary contact was very recent or, if ancient, that reproductive isolating mechanisms prevent their wholesale unification. Given their vagile nature and the lack of clear physical barriers separating them, we hypothesized that one or both lineages occupied smaller ranges removed from the current zone of contact throughout most of the Pleistocene and achieved contact only recently through a massive Holocene expansion. To investigate this hypothesis, we explored their demographic histories and population structure using a combination of whole-genome and reduced-representation sequencing. Additionally, we characterized the timing and extent of gene flow pulses between western and eastern gray foxes using a local ancestry inference-based approach.
In Chapter 2, we used both whole-genome (n = 26) and reduced representation (n = 197) sequencing to contrast the demographic histories of western and eastern gray foxes. Pairwise sequential Markovian coalescent (PSMC) modeling, stairway plots, and summary statistics of eastern and western gray foxes on either side of the contact zone showed contrasting demographic trajectories, with the trajectory of the eastern population declining and the trajectory of the western population increasing for most of their post-divergence history; during the latter portion of the last (Wisconsinan) glacial cycle and most of the Holocene, the eastern trajectory increased and the western trajectory declined. Correspondingly, the eastern lineage exhibited much lower genetic diversity than the western lineage, a cline in diversity consistent with a westward expansion front, and minimal genetic structuring. In contrast, the western lineage exhibited population structure and locally varying demographic histories, reflecting long-term occurrence over a broad region of the continent. The recurrent declines in the eastern population may have kept them both geographically and demographically restricted to the southeast for much of their evolutionary history, resulting in the deep divergence and limited gene flow currently observed between lineages. Additionally, population structure and variable demographic histories within the western lineages may reflect separation in distinct glacial refugia and subsequent gene flow across the western gray fox range.
In Chapter 3, I utilized whole genomes of gray foxes (n = 42) from both the western and eastern lineages as well as from the hybrid zone to investigate (1) the timing of secondary contact and genetic exchange, (2) the width of the hybrid zone in the context of this timing, and (3) signatures of selective introgression between lineages. I inferred the timing of admixture pulses using a local ancestry inference-based approach, which was optimized for low-coverage sequencing data. I tested whether observed patterns of admixture were consistent with expectations based on a model assuming no reproductive barriers. I then investigated specific genomic regions that were introgressed across the contact zone at unusually high frequencies, consistent with selective introgression. I identified two distinct pulses of late Holocene and historical admixture. The older pulse of admixture (3,500 YBP) reflected unidirectional gene flow from east to west, likely driven by a major demographic expansion of the eastern gray fox. In contrast, the more recent bi-directional pulse of admixture began approximately 200 YBP, coinciding with major anthropogenic landscape changes. Given the recency of genetic interchange, the narrow widths of the geographic clines provided little insight on the question of reproductive isolation but afforded an opportunity to explore selective introgression. Several genomic regions were identified as candidates for selective introgression and may have been associated with behavioral divergence, mate choice, and olfaction.
