Genetic architecture and accuracy of predicted genomic breeding values for sea lice resistance in the St John River aquaculture strain of North American Atlantic salmon

• Published in: Aquaculture 2024-05, Vol.586, p.740819, Article 740819
• Main Authors: Vallejo, Roger L., Pietrak, Michael R., Milligan, Melissa M., Gao, Guangtu, Tsuruta, Shogo, Fragomeni, Breno O., Long, Roseanna L., Peterson, Brian C., Palti, Yniv
• Format: Article
• Language: English
• Subjects: Atlantic salmon; Breeding value prediction; Genome-wide association study; Genomic selection; Lepeophtheirus salmonis; Sea lice resistance
• ISSN: 0044-8486; 1873-5622
• DOI: 10.1016/j.aquaculture.2024.740819

Sea lice infections cause significant economic losses in Atlantic salmon (Salmo salar) farming. The objectives of this study were to (1) estimate the heritability of resistance to sea lice in the USDA's breeding program for North American (N.A.) Atlantic salmon; (2) elucidate the genetic architecture of sea lice resistance; and (3) assess the accuracy and bias of predicted breeding values for sea lice resistance using cross-validation analysis (CVA) and progeny testing of selection candidates (PTSC). Fish from year-class (YC) 2014 n=967 and YC 2018 n=941 were challenged with sea lice (Lepeophtheirus salmonis) and genotyped with the new USDA 50K SNP chip developed for the N.A. Atlantic salmon. Heritability estimates for sea lice count in the two year-classes were between 0.07 and 0.10. GWAS with weighted ssGBLUP (wssGBLUP) highlighted the polygenic architecture of the trait. With CVA, the genomic prediction methods x¯=0.54 had higher accuracy of predictions than the pedigree-based method x¯=0.43; and ssGBLUP had the highest accuracy of BV predictions (0.50–0.62) among the genomic prediction methods. With PTSC approach, the pedigree-based method PBLUP had higher estimated accuracy than the genomic prediction methods (0.48 vs. 0.30). Bias of BV predictions with CVA x¯=0.87range0.57−1.06 was lower than with PTSC x¯=1.22range0.44−2.35. Our retrospective assessment of the accuracy of indirect sib-based selection to improve sea lice resistance in the St John River aquaculture stock that is commonly used for farming of Atlantic salmon in North America demonstrated that substantial genetic gains can be obtained. Due to the polygenic architecture of the trait, the estimated accuracy of the ssGBLUP model was better than the wssGBLUP model. In our assessment, a larger training sample size will be needed to achieve optimal results with the whole genome-enabled selective breeding method due to the polygenic architecture of sea lice resistance coupled with low heritability of the trait. •Polygenic architecture of sea lice resistance was confirmed in North American Atlantic salmon.•Substantial genetic gains for sea lice resistance can be achieved through indirect sib-based selection.•Larger training sample size will be needed to achieve optimal results with genome-enabled selective breeding method.•First report that evaluates the accuracy of genomic predictions for sea lice resistance using progeny performance data.