Genotype Imputation in Atlantic Salmon: Impact on Refinement of GWAS and Assessment of Accuracy for Cost-Effective Genomic Selection

Genotype imputation can increase power and refine results from genome-wide association studies (GWAs). Moreover, the practical implementation of low cost genomic selection in livestock is based on imputed genotypes from low to high density single nucleotide polymorphisms (SNP) panels. This study aimed at: i) determining the impact of imputation from medium-density (50K) to high-density (150K) genotypes on the refinement of association signals from GWAs for economically important traits in Atlantic salmon; and ii) determining the accuracy of imputed genotypes from low-density (0.5K and 6K) to medium-density (50K) SNP panels for the implementation of cost effective genomic selection schemes in this species. For the first objective, imputations from 50K to 150K were carried out with FImpute v.2.2 software using 53 parents genotyped with 150K and 4.691 offsprings genotyped with 50K. GWAS for body weight (BW), average daily gain (ADG) and resistance against Piscirickettsia salmonis infection (SRS) were performed using high density imputed genotypes, polygenic models and Family-based Score Test for Association methodology. For the second objective, five replicates were done for imputation from 0.5K and 6K to 50K, using a proportion of 10 and 90% of the animals in training and testing group, respectively. Genotypes imputed from 50K to 150K increased the power of GWAS analyses and identified novel and refined associations for BW, ADG and SRS. Accuracies of imputation from 0.5K and 6K up to 50K ranged from 0.81 to 0.98. Imputation strategies will allow the refinement of GWAs studies and the implementation of cost-effective genomic selection for Atlantic salmon.