Genomic predictions for daily gain and fillet weight using correlated size and body area measurements in Asian seabass (Lates calarifer, Bloch 1790).

Genomic prediction models have proven efficient for increasing genetic gains of growth and fillet traits in aquaculture. However, measurement of these traits is expensive, time-consuming, or cannot be performed on live animals. This study investigated the potential of using correlated body traits as an alternative basis for supporting genomic prediction of average daily gain (ADG) and fillet weight (FW) in Asian seabass. A total of 954 fish were genotyped for 20,430 single nucleotide polymorphisms (SNPs). Phenotypic data at 390 days, including body weight, fillet weight, 16 size and four body-part-area measurements were used. Heritability estimates were high for ADG (0.47) and FW (0.44), low to moderate (0.12–0.39) for body sizes and body areas. Genetic correlations among pairs of traits were all high and positive (0.65–0.99). Five prediction models (1, 2, 3, 4 and 5) were evaluated using the GBLUP method and were compared using five-fold cross validation. Results indicated that model 4, which fitted the secondary traits derived from correlated body sizes and body areas as the phenotypes, was the best model, showing prediction accuracy of 0.47 with small prediction bias (0.95). The results demonstrated the advantage of using correlated traits for practical genomic prediction in Asian seabass, particularly for growth and fillet traits that are difficult to measure. • Eight size and two body area traits were used as secondary traits for genomic prediction of ADG and FW. • Genomic prediction was performed on 954 fish genotyped at 20,430 SNPs using five GBLUP models. • Model 4, was the best model, showing prediction accuracy of 0.47 with small prediction bias (0.95). [ABSTRACT FROM AUTHOR]