The impact of selection, gene flow and demographic history on heterogeneous genomic divergence: three‐spine sticklebacks in divergent environments

Abstract

Heterogeneous genomic divergence between populations may reflect selection, but should also be seen in conjunction with gene flow and drift, particularly population bottlenecks. Marine and freshwater three-spine stickleback (Gasterosteus aculeatus) populations often exhibit different lateral armour plate morphs. Moreover, strikingly parallel genomic footprints across different marine-freshwater population pairs are interpreted as parallel evolution and gene reuse. Nevertheless, in some geographic regions like the North Sea and Baltic Sea, different patterns are observed. Freshwater populations in coastal regions are often dominated by marine morphs, suggesting that gene flow overwhelms selection, and genomic parallelism may also be less pronounced. We used RAD sequencing for analysing 28 888 SNPs in two marine and seven freshwater populations in Denmark, Europe. Freshwater populations represented a variety of environments: river populations accessible to gene flow from marine sticklebacks and large and small isolated lakes with and without fish predators. Sticklebacks in an accessible river environment showed minimal morphological and genomewide divergence from marine populations, supporting the hypothesis of gene flow overriding selection. Allele frequency spectra suggested bottlenecks in all freshwater populations, and particularly two small lake populations. However, genomic footprints ascribed to selection could nevertheless be identified. No genomic regions were consistent freshwater-marine outliers, and parallelism was much lower than in other comparable studies. Two genomic regions previously described to be under divergent selection in freshwater and marine populations were outliers between different freshwater populations. We ascribe these patterns to stronger environmental heterogeneity among freshwater populations in our study as compared to most other studies, although the demographic history involving bottlenecks should also be considered in the interpretation of results.

Cryptic invasion drives phenotypic changes in central European threespine stickleback

“…I counted all lateral plates, including the keel plates, on the left flank for both historical and contemporary specimens from Basel and Village-Neuf. Additional lateral plate phenotypes from the same time period were available for 40 individuals from the Basel site (Fatio 1882; Table S1 in supplementary material). I compared the number of lateral plates among contemporary and historic (including additional phenotypes from Fatio) samples using KruskalWallis rank sum tests because plate numbers were not normally distributed (Fig.…”

“…Freshwater stickleback commonly derive from ancestral marine populations that colonized freshwater habitats since the last glacial period *12,000 years ago (McKinnon and Rundle 2002). In the River Rhine, stickleback were naturally restricted to as far upstream as Basel and were absent from the Swiss midlands until about 1870 (Fatio 1882;Lucek et al 2010). Following their subsequent introduction, and facilitated by the channelization of many Swiss waterways, stickleback have since invaded a large part of the country.…”

“…In contrast, individuals from the upper Rhone River and the Lake Geneva region have only few lateral plates and derive from a lineage found in the lower Rhone (Lucek et al 2010). Whereas the aforementioned lineages were introduced in the 1870s, a third and presumably native lineage with few lateral plates, dominates the River Rhine south of Basel (Fatio 1882;Lucek et al 2010). These distinct lateral plate phenotypes have a simple genetic architecture that is linked to the Ectodysplasin (Eda) locus (Colosimo et al 2005).…”

“…Historic records for central Europe (Bertin 1925;Münzing 1963;Gross 1977) and particularly the Basel region (Fatio 1882;Steinmann 1936; Table S1 in supplementary material; see Fig. 1), representing the southernmost part of the native range of stickleback in the River Rhine, suggest that these populations were fixed for a phenotype with only few lateral plates.…”

Exploring the phylogeography of a hexaploid freshwater fish by RAD sequencing

“…This method is popular and has been used in many studies since its inception (Figure S1). RAD sequencing has been used, particularly in fish, to identify population divergence (Boehm, Waldman, Robinson, & Hickerson, 2015; Ferchaud & Hansen, 2016; Larson et al., 2014), for SNP identification in polyploid fish (Hohenlohe, Amish, Catchen, Allendorf, & Luikart, 2011; Ogden et al., 2013; Palti et al., 2014), in phylogeographic studies (Macher et al., 2015; Reitzel, Herrera, Layden, Martindale, & Shank, 2013), for QTL analysis (Gagnaire, Normandeau, Pavey, & Bernatchez, 2013; Houston et al., 2012; Yoshizawa et al., 2015), for linkage mapping (Brieuc, Waters, Seeb, & Naish, 2014; Henning, Lee, Franchini, & Meyer, 2014), in hybridization studies (Hand et al., 2015; Lamer et al., 2014; Pujolar et al., 2014), for exploration of genome architecture and evolution (Brawand et al., 2014; Kai et al., 2014; Waples, Seeb, & Seeb, 2016), and in phylogenetic analyses (Gonen, Bishop, & Houston, 2015; Wagner et al., 2013). This methodology should be particularly suited to phylogeographic studies as the inference power from large numbers of markers may identify patterns that are not easily visible in traditional analyses based on relatively few loci (Davey et al., 2011).…”