Northern European Salmo trutta (L.) populations are genetically divergent across geographical regions and environmental gradients

Abstract

The salmonid fish Brown trout is iconic as a model for the application of conservation genetics to understand and manage local interspecific variation. However, there is still scant information about relationships between local and large-scale population structure, and to what extent geographical and environmental variables are associated with barriers to gene flow. We used information from 3,782 mapped SNPs developed for the present study and conducted outlier tests and gene-environment association (GEA) analyses in order to examine drivers of population structure.Analyses comprised >2,600 fish from 72 riverine populations spanning a central part of the species' distribution in northern Europe. We report hitherto unidentified genetic breaks in population structure, indicating strong barriers to gene flow. GEA loci were widely spread across genomic regions and showed correlations with climatic, abiotic and geographical parameters. In some cases, individual loci showed consistent GEA across the geographical regions Britain, Europe and Scandinavia. In other cases, correlations were observed only within a sub-set of regions, suggesting that locusspecific variation was associated with local processes. A paired-population sampling design allowed us to evaluate sampling effects on detection of outlier loci and GEA.Two widely applied methods for outlier detection (pcadapt and bayescan) showed low overlap in loci identified as statistical outliers across sub-sets of data. Two GEA analytical approaches (LFMM and RDA) showed good correspondence concerning loci associated with specific variables, but LFMM identified five times more statistically significant associations than RDA. Our results emphasize the importance of carefully considering the statistical methods applied for the hypotheses being tested in outlier analysis. Sampling design may have lower impact on results if the objective is to identify GEA loci and their population distribution. Our study provides new insights into trout populations, and results have direct management implications in serving as a tool for identification of conservation units. | 401BEKKEVOLD Et aL.

Genomic Hatchery Introgression in Brown Trout (Salmo trutta L.): Development of a Diagnostic SNP Panel for Monitoring the Impacted Mediterranean Rivers

“…In our study, using a 2b-RAD method for the first time for high-throughput SNP genotyping, 24,830 nuclear SNPs, distributed among the 40 assembled chromosomes of the brown trout genome, were retained after applying all quality and population filters. The SNP density obtained was within the broad range of other brown trout genome-wide studies, from ~3 K [ 94 , 95 ] to ~100 K [ 52 , 96 ]. Further, this study is a pioneering investigation, using the brown trout reference genome for genotyping (see [ 97 ]) due to its recent availability (NCBI June 2019; [ 55 ]).…”

Genomic data support management of anadromous Arctic Char fisheries in Nunavik by highlighting neutral and putatively adaptive genetic variation

“…We explored how genetic variation in Arctic Char was linked to a range of climatic and physical environmental predictors while considering both freshwater and marine habitats, something that has rarely been done in salmonids (but see Bekkevold et al, 2020 ). While different GEA methods identified candidate SNPs with every environmental component considered, we found that most of the best‐supported candidates were associated with the component reflecting summer SST and air temperature, salinity, and turbidity (PC2).…”

Spotting genome-wide pigmentation variation in a brown trout admixture context

“…Jombart, Pontier and Dufour (2009) already mentioned that RDA was somewhat neglected in association studies despite now recognized desirable properties to limit false positives (Capblanc, Luu, Blum, & Bazin, 2018; Forester et al, 2018) and robustness to recombination rate variation (Lotterhos, 2019). Effectively, the use of RDA to investigate genotype-phenotype associations remains seldom (Talbot et al, 2017; Vangestel, Eckert, Wegrzyn, St. Clair, & Neale, 2018; Carvalho et al, 2020), while it became a standard in genotype-environment association studies (Forester, Lasky, Wagner, & Urban, 2018), including trout (Bekkevold et al, 2019). Basically, RDA allowed for distinguishing the part of variation collectively explained by RAD-loci and independent pigmentation variables, and hereby estimated to ∼35% of observed trait variation.…”