Introduction:
Genome scans are useful tools for identifying the effects of evolutionary processes on the genome of a species (Lotterhos and Whitlock, 2015; Fraser and Whiting, 2019). In the past decade they have been used to analyse genomic patterns in many wild species (Alves et al., 2019; Dennenmoser, Vamosi, Nolte, & Rogers, 2017; Jones et al., 2012; Vijay et al., 2016; Westram et al., 2014), as they can provide genetic information about evolution without requiring typically impractical experimental setups. The growth of studies using genome scans has provided a new opportunity to compare results among species to identify common patterns of genetic variation, which may be imprinted on different species through the same evolutionary processes. Ultimately, comparisons of genome scans among species will help to assess the generality of genetic patterns to learn how evolution shapes the genomes of different species.
At the simplest level, genome scans are a comparison of genetic diversity among different populations within a species. Genetic diversity can be split into two main types; diversity within a population and diversity among populations (referred to as genetic divergence). Many statistics represent genetic diversity (e.g. π, HE, Tajima’s D, and Fay & Wu’s H) or genetic divergence (e.g. FST, dxy), and different interpretations of these scores have been discussed at length in other papers (Burri et al., 2015; Ellegren et al., 2012; Reid et al., 2016; Van Doren et al., 2017; Vijay et al., 2016, 2017). A genome scan moves along the genome looking for extreme patterns of these statistics that may be associated with local adaptation (Lotterhos & Whitlock, 2015, Fraser and Whiting 2019), but alternatively could be the product of background selection (Charlesworth et al., 1993, Matthey-Doret & Whitlock, 2019) or demographic events such as range expansions, population bottlenecks or inbreeding (Barton, 1998; Excoffier & Ray 2008; Lotterhos & Whitlock, 2014; Nielsen, Hellmann, Hubisz, Bustamante, & Clark, 2007). These extreme patterns can be identified visually as “peaks” and “troughs” of genetic diversity or divergence, from their distinctive shape on a Manhattan plot. Statistical methods are used to determine which evolutionary processes most likely generated these peaks and troughs, often as the first step towards identifying candidate genes.
Comparison of genome scan results among species provides insight into how shared ancestry, demography, and environmental conditions can affect the similarity of patterns in their genomes. Commonly, genome scans are compared to detect convergent evolution (Fraser and Whiting 2019), as shared peaks or troughs have the potential to reveal genes that underpin evolution to a shared environmental pressure in many species (Stern 2013). Examples of these convergently evolving genes have already been found such as digestive proteins in primates (Stewart, Schilling, & Wilson, 1987), pigmentation in vertebrates (Gompel & Prud’homme, 2009; Hoekstra, 2006; Manceau et al., 2010) or anthocyanin proteins in flowering plants (Kopp 2009). Outside of convergent evolution, comparing genome scans can also show shared properties of the genome such as recombination landscapes (Samuk et al. , 2017) or ancestral population structure (Vijay et al. , 2017). On one hand, genomes scans should not be used in isolation to detect convergent evolution, as shared patterns can come from several sources. On the other hand, genome scans offer a useful way to identify broad scale genetic similarities among several species. By comparing patterns in diversity and divergence across many species and environmental gradients, we can better understand how evolutionary processes affect the genome.
Threespine stickleback (Gasterosteus aculateus ) is a good system for comparative genome scans, as several regions of the genome have been identified that are strongly associated with local adaptation in this species (Colosimo et al., 2005; Hohenlohe et al., 2010; Jones et al., 2012; Schluter & Conte, 2009). Several closely related fish species live in overlapping niches allowing their genomic landscape to be compared to the threespine stickleback’s to learn how evolution shapes patterns in their respective genomes. This study aims to compare patterns of genetic diversity and divergence in the threespine stickleback with both the ninespine stickleback (Pungitus pungitus ; for simplicity the stickleback species will be referred to as threespines and ninespines) and tubesnout (Alurhychous flavidus ), as an example of how comparisons of genome scan results can identify common genetic patterns.
Ninespines and threespines diverged 26mya (Varadharajan et al., n.d.) and have already been subjected to comparative genetic studies (Varadharajan et al. , no date; Shikano et al. , 2013; Nelson and Cresko, 2018), in part because both species have colonised freshwater lakes in similar regions. Interestingly, while targeted genetic studies support convergent evolution to freshwater (Shikanoet al. , 2013), whole genome data found no genetic signatures of convergent evolution (Raeymaekers et al. , 2017). The extent of similarity in genetic patterns among these sticklebacks is still an open question.
We are only beginning to compare the genomes of the threespine and tubesnout (Li et al. in review) and have yet to explore the patterns of genetic diversity. These species diverged approximately 50mya (Betancuret al. , 2013), which is a timeframe similar to a study in birds which found similar patterns of genetic diversity maintained across 55 million years (Vijay et al. , 2017). In contrast to the ninespine-threespine comparison, tubesnouts are an exclusively marine species that overlaps with the marine threespine along most of its range in the Pacific. Marine threespines are known to have genetic structure along the North American West coast (Morris et al. , 2018), which may be the result of gene flow from locally-adapted freshwater populations (Nelson and Cresko, 2018). Thus, we may expect to find patterns in the threespine genome that differ from the tubesnout’s, due to differences in their demographic history, selection, and ancestral variation.
Here, we compare patterns of population genomic diversity and divergence in these species to assess how such patterns vary across the stickleback order. Specifically, we study patterns in FST and genetic diversity from populations at each end of a latitudinal gradient and compare these patterns among species-pairs at a whole-genome and a gene-by-gene level to assess their similarity and test for signatures of convergent evolution. We focus on latitude-related effects (e.g. adaptation in traits related to body size, growth rate, changing breeding times or oxygen binding [Andersen et al., 2009; Bell & Foster, 1994, pp. 155–157; Blanck & Lamouroux, 2007]) instead of the patterns of salinity-driven adaptation more commonly investigated in threespine and ninespine, as the tubesnout has not evolved to live in freshwater systems. By studying broad-scale patterns that covary with the selection pressures associated with latitude, we aim to detect whether patterns of genetic diversity are shared among these species, to learn how evolution may have shaped such patterns.