Species delimitation and dispersal capabilities within the C. variipennis complex
The high degree of genetic differentiation between clusters inferred by the SNP data supports the current species groupings of the C. variipennis complex (C. occidentalis , C. sonorensis , andC. variipennis ), as well as raising C. albertensis and a cryptic species in San Diego, California to species status. Little to no IBD or structure was found within populations of C. albertensis ,C. sonorensis , and C. variipennis (Fig. 3a,c,d). The number of populations inferred by fastSTRUCTURE for C. sonorensiswas K=2; however, a mean pairwise FST of 0.0287 suggests that a high amount of gene flow still exists between all populations. This could also be an artifact of the propensity of delta K inferring two populations (Janes et al., 2017) or from a high level of relatedness among individuals from KS.
Interestingly, although no IBD was found in C. occidentalis , each location of this species clustered as a distinct population. The lack of IBD is therefore not indicative of a single, genetically homogeneous population, but rather stems from high levels of divergence between populations regardless of their geographic distances. In this species, the strong genetic divergence between the population from California and the other populations observed in the SNP data was consistently uncovered in the mtDNA (4.0% divergent, Table 3, Fig. 5). It is possible that this may represent a further cryptic species with a dispersal barrier created by the Sierra-Nevada mountain range. Patchiness of the larval habitat of C. occidentalis could also create isolation between populations as well as reduce the number of individuals within each population. A small population size with little to no immigration would allow for a strong effect from drift (Hare, 2001). While the populations of C. occidentalis outside of California were less diverged from one other, the lowest pairwiseFST values between these populations were still greater than the highest pairwise values observed for any other species, consistent with the findings of Holbrook et al. (2000) (Table 2). Interestingly, at one of the three loci found to be under selection with the complex (seipin, Table S4), all populations of C. occidentalis and C. albertensis were fixed for a single allele, whereas the other three other species were fixed for the other alternative allele. This SNP was determined to be synonymous and therefore unlikely to be the direct target of selection; however, it may be linked to a region of the genome that is.
Similar to other species of Culicoides (Jacquet et al., 2015; Mignotte et al., 2021; Onyango, Beebe, et al., 2015; Onyango, Michuki, et al., 2015), high values of the inbreeding coefficient were observed in all species investigated in this study (Table S2). Although these previous studies have suggested that the observed high inbreeding coefficient values are an artifact from a large number of null alleles, the consistent reporting of these findings across various species using several types of molecular markers lends support to the they hypothesis that high inbreeding has a biological origin. High levels of inbreeding and heterozygote deficiencies are common among mosquitoes (Fonseca, Smith, Kim, & Mogi, 2009; Goubert, Minard, Vieira, & Boulesteix, 2016; Lehmann et al., 1997), even when using markers with a low level of null alleles (Chapuis & Estoup, 2006; Manni et al., 2015). Goubert et al . (2016) considered the typical Aedes albopictus population as “a network of interconnected breeding sites, each with a high level of inbreeding”. In this study, although we cannot rule out that the presence of null alleles and we acknowledge that a weak Wahlund effect can contribute to the level of inbreeding, our results strongly suggested that some aspects of the reproductive biology ofCulicoides induce inbreeding within populations.