Introduction
Speciation is a dynamic evolutionary process through which populations segregate into independently evolving lineages over time (De Queiroz, 2007). When gene flow is restricted between populations, the accumulation of genetic changes, through selection or local genetic drift, may lead to genetic differentiation and potentially reproductive isolation (Coyne & Orr, 2004; Endler, 1973; Mayr, 1999; Richardson, Urban, Bolnick, & Skelly, 2014). This restriction of gene flow occurs through either geographic or ecological isolation, though these are not mutually exclusive (Nosil, 2008). Thus, the level of gene flow between divergent populations is a contributing factor to the rate of speciation, as well as to the spatial level at which it occurs (Kisel & Barraclough, 2010).
Geographic isolation reduces migration between populations, and thus, life-history traits influencing dispersal ability can drastically influence the level of gene flow among populations (e.g. (Bolnick & Otto, 2013; Slatkin, 1993). Species with low dispersal ability are particularly likely to exhibit highly differentiated populations resulting in the evolution of cryptic species over a limited spatial scale (Pante, Puillandre, et al., 2015). In contrast, species with high dispersal abilities are likely to maintain gene flow between populations, therefore a process outside of geographic isolation is needed to initiate the speciation process in these instances (Claramunt, Derryberry, Remsen Jr, & Brumfield, 2012; Enbody et al., 2021; Palumbi, 1994). Two such mechanisms for this are ecological and behavioral isolation whereby sympatric populations occupy distinct ecological niches (Rundle & Nosil, 2005; Wang & Bradburd, 2014). The difference between these populations, such as habitat type, differences in mating times, or sexual selection reduce the frequency of interbreeding and this can lead to genetic divergence (Coyne & Orr, 2004; Dobzhansky, 1982). Ecological and behavioral isolation can drive lineage divergence through selection, and subsequent pre-zygotic isolation can further increase divergence through reinforcement, accentuating the speciation process (Coyne & Orr, 2004). However, with an increase in specialization, fragmented distributions of either habitat or host may further reduce gene flow between populations (Thompson, 1999; Thompson & Cunningham, 2002; West & Herre, 1994).
Each reproductive isolation mechanism can lead to similar morphological adaptations and genomic signatures (Keller et al., 2013; Seehausen et al., 2014), and this can make it challenging to interpret what factors contributed to the speciation event. While the most accurate assumptions about species delimitation are derived from a multifaceted approach (Carstens, Pelletier, Reid, & Satler, 2013; Schlick-Steiner et al., 2010), gene flow is directly tied to the fate of incipient species. By using a population genetic approach to study microevolution and incipient speciation, we can identify independent lineages and measure the introgression between them to better understand the evolutionary processes underlying species divergence. Species delimitation is especially important when working with organisms responsible for pathogen transmission, as misidentifications will lead to inaccurate vector competency and surveillance data. Culicoides Latreille (Diptera: Ceratopogonidae) biting midges are responsible for the transmission of many disease-causing agents worldwide (Borkent, 2004; Mellor, Boorman, & Baylis, 2000), including bluetongue virus (BTV) and epizootic hemorrhagic disease virus (EHDV). These viruses can cause severe symptoms and death in wild and domestic ungulates and are responsible for substantial economic losses globally (Rushton & Lyons, 2015; Tabachnick, 1996).
In North America, one of the main BTV and EHDV vectors is Culicoides sonorensis Wirth and Jones, which belongs to the C. variipennis species complex. When originally described, this group consisted of five subspecies (Wirth & Jones, 1957), but it is currently composed of three distinct species (C. occidentalis Wirth and Jones, C. sonorensis , andC. variipennis (Coquillet)) (Holbrook et al., 2000). Despite the current taxonomic arrangement, species identification remains difficult due to very subtle adult morphological differences and genetic similarity. Additionally, cryptic species could make vector incrimination and species distribution records potentially unreliable. Measuring genetic divergence between species and populations can be useful in vector biology as vectorial capacity and host association become increasingly variable with increased genetic distance (Byrne & Nichols, 1999; McCoy, Boulinier, Tirard, & Michalakis, 2001). Population genetic studies of Culicoides species in Europe, Africa, and Australia have consistently revealed frequent gene flow between populations, even at continental scales (Jacquet et al., 2015; Jacquet et al., 2016; Onyango et al., 2016; Onyango, Michuki, et al., 2015). Their high dispersal ability, likely wind-mediated (Ducheyne et al., 2007; Purse et al., 2005), decreases the likelihood of geographic isolation between populations of Culicoides spp. Under laboratory conditions, the species within the C. variipennis complex have been shown to hybridize (Velten & Mullens, 1997), and while C. occidentalis and C. variipennis are not known to be competent vectors, both species occur sympatrically with C. sonorensis(Wirth & Jones, 1957). This lack of post-zygotic reproductive isolation, coupled with a high dispersal ability and numerous sympatric populations, makes this species complex an intriguing system in which to study speciation and may also provide insights into the evolutionary mechanisms responsible for vector competence in this group.
Here, we evaluated the geographic connectivity within and among the species of the C. variipennis complex by assessing the level of gene flow within and across populations. We used a high-throughput ddRadSeq protocol to analyze 206 individuals collected from 17 sites throughout the United States and Canada. We first estimated the overall genetic similarity and population structure among these samples to determine distinct lineages within the species complex. We then estimated the level of gene flow between the inferred species, as well as uncovered hybridization between sympatric species. As previous attempts to separate these species using common barcoding genes have been inconclusive, we sequenced a region of COI to compare to the inferred SNP identifications. One species, was found to have two distinct geographic haplogroups, while three other species shared a single haplogroup. Additionally, we assessed the potential drivers of divergence in this species complex by assessing loci under selection for each species, as well as discuss the potential mechanisms controlling reproductive isolation.