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

Abstract

Colour and pigmentation variation attracted fish biologists for a while, but high-throughput genomic studies investigating the molecular basis of body pigmentation remain still limited to few species and conservation biology issues ignored. Using 75,684 SNPs, we investigated the genomic basis of pigmentation pattern variation among individuals of the Atlantic and Mediterranean clades of the brown trout (Salmo trutta), a polytypic species in which Atlantic hatchery individuals are commonly used to supplement local wild populations. Using redundancy analyses and genome-wide association studies, a set of 384 “colour patterning loci” (CPL) significantly correlated with pigmentation traits such as the number of red and black spots on flanks, but also the presence of a large black stain on the opercular bone was identified. CPLs map onto 35 out of 40 brown trout linkage groups indicating a polygenic basis to pigmentation patterns. They are mostly located in coding regions (43.4%) of 223 candidate genes, and correspond to GO-terms known to be involved in pigmentation (e.g. calcium and ion-binding, cell adhesion). Annotated genes especially include 24 candidates with known pigmentation effects (e.g. SOX10, PEML, SLC45A2), but also the Gap-junction Δ2 (GJD2) gene that was previously showed be differentially expressed in trout skin. Patterns of admixture were found significantly distinct when using either the full SNP data set or the set of CPLs, indicating that pigmentation patterns accessible to practitioners are not a reliable proxy of genome-wide admixture. Consequences for management are discussed.

Identifying and conserving sympatric diversity in trout of the genus Salmo, with particular reference to Lough Melvin, Ireland

“…F 1 offspring of gillaroo and sonaghen reared in separate small lakes s c150 km from Lough Melvin demonstrate the stability of their pigmentations patterns even under distinct environmental and feeding conditions (Figure 1). Genomic data are increasingly demonstrating the genetic bases of colouration differences in Eurasian trout (Valette et al, 2020). Stepwise discriminant analysis based on 12 meristic counts and 19 morphometric measurements from the three morphs correctly assigned 100% of the ferox, 98.8% of gillaroo and 98.7% of sonaghen (Cawdery & Ferguson, 1988).…”

Phenotypic Colour Variants in Batches of Fingerlings of Clarias gariepinus from Some Commercial Fish Hatcheries

“…Knowledge of the genetics that underpin the manifestation of various chromatypes or phenotypes of skin pigmentation has been based on qualitative traits through studies on natural or artificially activated colour mutants in species of fish that are used as models for study including zebrafish, Danio rerio [6] and medaka, Oryzias latipes [7]. Studies have shown that the genetic basis of skin pigmentation in fish can be any of monogenetic [8], polygenetic [9], pleiotropic and sex-linked [10,11]. The evidence from these studies overwhelmingly supports the involvement of several genes in the manifestation of a particular skin colour or colour pattern through modulation of chromatophores that lead to synthesis and expression of pigments.…”

Adaptive phenotypic and genetic variation in chickens: a landscape genomics approach

“…The use of RDA in genomewide association studies (GWAS) in animal breeding is not common. However, RDA is a powerful approach that can be utilized in GWAS studies in livestock (Kess and Boulding, 2019;Torrado et al, 2020;Valette et al, 2020). Phenotype association analysis was performed with RDA to explain variation in the genome associated with five quantitative traits (mature live body weight, beak length, comb width, wattle width and earlobe width).…”

“…Signatures of selection analysis identify regions of the genome that have differentiated between populations, possibly in response to selective pressure (Sabeti et al, 2006;Voight et al, 2006). Multivariate methods that simultaneously account for multiple drivers of phenotypic and environmental divergence, are recently being applied in landscape genomic studies to identify quantitative trait loci (QTL) associated with environment predictors (Forester et al, 2018;Harrisson et al, 2017;Kess and Boulding, 2019;Torrado et al, 2020) and with phenotypic variables (Carvalho et al, 2021;Kess and Boulding, 2019;Talbot et al, 2017;Valette et al, 2020;Vangestel et al, 2018). Multivariate ordination methods such as RDA have outperformed mixed-modelbased methods and machine learning-based methods (e.g., Random Forest) in detecting loci associated with environmental variation (Capblancq et al, 2018;Forester et al, 2018).…”

“…Multivariate ordination methods such as RDA have outperformed mixed-modelbased methods and machine learning-based methods (e.g., Random Forest) in detecting loci associated with environmental variation (Capblancq et al, 2018;Forester et al, 2018). Despite its ability to investigate genotype-phenotype associations RDA is mostly neglected in GWAS studies, while it became a standard in genotype-environment association studies (Jombart et al, 2009;Valette et al, 2020). We implemented a robust sampling strategy, considering environmental gradation (e.g., elevational clines) and geographic (latitudinal and longitudinal) variation in the country .…”