Population Structure Limits Parallel Evolution in Sticklebacks

Abstract

Population genetic theory predicts that small effective population sizes (Ne) and restricted gene flow limit the potential for local adaptation. In particular, the probability of evolving similar phenotypes based on shared genetic mechanisms (i.e. parallel evolution), is expected to be reduced. We tested these predictions in a comparative genomic study of two ecologically similar and geographically co-distributed stickleback species (viz. Gasterosteus aculeatus and Pungitius pungitius). We found that P. pungitius harbours less genetic diversity and exhibits higher levels of genetic differentiation and isolation-by-distance than G. aculeatus. Conversely, G. aculeatus exhibits a stronger degree of genetic parallelism across freshwater populations than P. pungitius: 2996 vs. 379 SNPs located within 26 vs. nine genomic regions show evidence of selection in multiple freshwater populations of G. aculeatus and P. pungitius, respectively. Most regions involved in parallel evolution in G. aculeatus showed increased levels of divergence, suggestive of selection on ancient haplotypes. In contrast, haplotypes involved in freshwater adaptation in P. pungitius were younger, and often associated with reduced diversity. In accordance with theory, the results suggest that connectivity and genetic drift play crucial roles in determining the levels and geographic distribution of standing genetic variation, providing evidence that population subdivision limits local adaptation and therefore also the likelihood of parallel evolution.

AF-vapeR: A multivariate genome scan for detecting parallel evolution using allele frequency change vectors

“…We applied SNP across most regions at these extended physical sizes. With 10 SNP windows (N=2376), the genomic landscape of full parallelism, defined as the per-window primary eigenvalues (Figure 4B), closely resembled the genomic landscape of marine-freshwater divergence commonly observed in this system (Fang et al, 2021;Hohenlohe et al, 2010;Roberts Kingman et al, 2021). In particular, chromosome 4 exhibited a clear chromosome-level increase in parallelism (Figure 4C-D).…”

“…This involves sequencing populations in replicated environments, performing genome scans for signatures of selection, and comparing candidate loci across replicates (reviewed in (Fraser & Whiting, 2019)). Once selection candidates have been identified, they can be probed for theories of genetic convergence, for example their mode of convergence (Lee & Coop, 2017), the role of genetic and pleiotropic constraints (Morris et al, 2019;Wollenberg Valero, 2020), the influence of redundancy in the genotype-phenotype map (Láruson et al, 2020;Yeaman et al, 2018), or the effect of gene flow (Fang et al, 2021). It is also of interest whether repeatability occurs through: 1) repeated selection at the same SNP (often termed 'parallel' evolution) (e.g.…”

Cast away in the Adriatic: Low degree of parallel genetic differentiation in three‐spined sticklebacks

“…This scenario of gene flow between marine and freshwater populations is also the underlying cause of why threespined stickleback populations frequently exhibit high levels of genetic parallelism in response to freshwater colonization (Barrett & Schluter, 2008;Schluter & Conte, 2009). However, among global G. aculeatus populations the degree of parallel genetic changes that result in similar phenotypes through shared molecular evolutionary pathways (Arendt & Reznick, 2008) might not be as high or as geographically uniform as previously assumed (Fang et al, 2021;Ferchaud & Hansen, 2016;Liu et al, 2018;Terekhanova et al, 2014).…”

“…The explanation for this appears to be varying levels and access to SGV for freshwater adaptation in different parts of the world (Fang et al, , 2021Kemppainen et al, 2021). Pools of SGV are significantly influenced by the ancestral (marine) populations' N e , which is the single most important determinant of genetic drift in a population (MacPherson & Nuismer, 2017;Thompson et al, 2019), as well as by population connectivity, which enables access to the same freshwater-adapted alleles across wider geographical areas (Bailey et al, 2017;Lee & Coop, 2017;Ralph & Coop, 2015;Terekhanova et al, 2014).…”

“…To detect genomic regions involved in parallel evolution associated with marine-freshwater differentiation in the Adriatic Sea and elsewhere in the world, we used linkage disequilibrium network analyses (Kemppainen et al, 2015;Li et al, 2018) together with linear mixed models (Fang et al, 2021). In addition, given that population history is an important determinant of the likelihood of parallel evolution (Conte et al, 2012), we reconstructed demographic histories of the Adriatic three-spined stickleback populations to determine their divergence times, population size changes and connectivity.…”