Parallel adaptive changes under replicated environmental conditions have been particularly valuable for understanding evolutionary processes in natural populations. One of the classical questions in evolutionary biology concerns whether different species and populations within species will adapt to the same agent of selection in the same way or whether the response will involve different traits and genes [1, 2]. Parallel genotypic adaptation appears to be frequent and occurs at all taxonomic levels from microbes and plants to humans [3, 4] and is likely to result in changes at a relatively small number of genes . For instance, the study of Colosimo et al.  demonstrated that selection on a single gene, ectodysplasin (Eda), is responsible for the parallel reduction of armor plates in freshwater populations of threespine stickleback Gasterosteus aculeatus. However, more complex physiological processes relevant in the context of parallel freshwater adaptation of threespine sticklebacks are influenced by several genes, each of small effect [7–10]. Using a survey of the published literature on parallel adaptation of independent lineages of natural populations, Conte et al.  concluded that divergence at loci under selection is most likely to be based on standing genetic variation derived from a common ancestor rather than mutations occurring de novo after divergence. Hence, probability of gene reuse is plausibly higher in closely related species, which are likely to show similar divergence at loci subjected to similar selection pressures .
An excellent opportunity to test for genetic parallelism exists in the two North Atlantic eel species, the European eel (Anguilla anguilla) and the American eel (A. rostrata). Both species are morphologically almost indistinguishable, with the number of vertebrae being regarded as the best diagnostic character between species . Divergence time between the two species remains largely unresolved, encompassing between 1.5 and 5.8 million years [13–15]. Remarkably, although mitochondrial DNA lineages of the two species are reciprocally monophyletic , differentiation at nuclear loci is surprisingly low (FST = 0.055 ; FST = 0.018 ; FST = 0.06 ; FST = 0.09 ), suggestive of ongoing gene flow. In this sense, it is well established that the spawning grounds of the two species overlap in the Sargasso Sea and there is also overlap in spawning time . European and American eels are known to hybridize, with hybrids observed almost exclusively in Iceland [21–23]. Hence, the sister species status of European and American eel and the low but biologically significant gene flow makes them an adequate system in which to test the occurrence of selection at homologous loci within each species.
North Atlantic eels have a catadromous life cycle and after spawning in the Sargasso Sea, larvae are transported by the Gulf Stream and other currents to the shores of North America and Europe/North Africa, respectively. Upon reaching the continental shelf, larvae metamorphose into glass eels, which complete the migration into riverine, estuarine and coastal feeding habitats and grow up as yellow eels. After a highly variable feeding stage, yellow eels metamorphose into partially mature silver eels that migrate back to the Sargasso undertaking a journey of about 2,000 km for the American eel and 5,000-6,000 km for the European eel. Upon arriving in the Sargasso Sea, eels reproduce and die . During the continental phase, eels occupy a broad range of habitats from the Caribbean to Greenland in the western Atlantic (American eel) and from Morocco to Iceland in the eastern Atlantic (European eel). The presence of eels in extremely heterogenous environments in terms of temperature (i.e. from subtropical to subarctic), salinity (i.e. from freshwater to marine), substrate, depth or productivity along their geographic distribution makes them ideal species in which to study the consequences of spatially varying selective pressures that often result in local adaptation of ecologically important traits [1, 25, 26]. Beginning with Levene , who introduced the first theoretical model for examining the impact of diversifying selection in space, a number of studies have shown that balancing selection due to spatial heterogeneity is an important mechanism responsible for the maintenance of genetic polymorphism (reviewed in ). Genetic variation in a spatially heterogenous environment may be maintained even when dispersal results in complete mixing of the gene pool . However, under such a panmixia scenario, in which offspring are distributed to environments at random independently of the environment experienced by the parents, local selection cannot to lead to local adaptation . In the case of eels, owing to panmixia in both European [19, 30] and American eel  and random larval dispersal across habitats, heritable trans-generational local adaptation is not possible although single-generation footprints of selection can still be detected. In this sense, significant geographic clines at allozyme loci have been detected in both European  and American eel . In the most comprehensive study to date, Gagnaire et al.  found evidence for spatially varying selection at 13 coding genes in American eel showing significant correlations between allele frequencies and environmental variables (latitude, longitude and temperature) across the entire species range.
In this study, we tested for single-generation signatures of spatial varying selection in European eel and compared the results to those obtained by Gagnaire et al. . We genotyped glass eels from 8 sampling locations across the geographic distribution of the species, using the same set of SNPs analyzed by Gagnaire et al.  in American eel. We used two main analytical approaches, one that identifies outliers as those markers with greater differentiation among all SNPs and a second based on determining positive associations between allele frequencies and environmental factors. Following the positive associations observed by Gagnaire et al.  in American eel, variables used in our study were degrees North latitude, degrees East/West longitude and sea-surface temperature at river mouth. We specifically wanted to test whether the same genes were under spatially varying selection in both European and American eel, hence providing evidence for parallel patterns of local selection, or whether the response involved a different set of genes. Considering their sister species status and the existence of gene flow between species, together with the similar environmental conditions they encounter , we hypothesize that the two North Atlantic eel species show parallel patterns of selection at the same loci.