Introduction and Methodology
Olive flounder (Paralichthys olivaceus) is the dominant marine aquaculture species in South Korea, yet its production system faces increasing challenges due to reliance on moisture pellets (MP), which depend on limited fishery resources and pose sustainability concerns. Although extruded pellets (EP) provide practical and environmental advantages, their widespread adoption has been hindered by reduced growth performance. Improving growth under EP-based feeding systems is therefore a critical objective for sustainable aquaculture. In this study, we investigated the genetic architecture of growth traits and evaluated genomic prediction strategies under two EP-based diets differing in fishmeal content: a commercial diet and a low-fishmeal diet. A large-scale population was established, and individuals were repeatedly phenotyped for growth traits, including body weight (BW), weight gain rate (WGR), and specific growth rate (SGR), and genotyped using a high-density SNP array.
Results and Discussion
Genome-wide association analysis revealed that growth traits were predominantly polygenic under both dietary conditions, with only a single SNP reaching genome-wide significance under the low-fishmeal diet. Heritability estimates indicated that BW consistently exhibited moderate to high heritability, whereas WGR and SGR showed relatively lower values. These results highlight BW as the most reliable trait for genomic improvement. Genomic prediction analyses demonstrated that predictive performance varied across traits and diet groups, with BW achieving the highest accuracy. Notably, incorporating informative SNPs identified through GWAS markedly improved prediction accuracy, with values approaching 0.8. Evaluation of genotype-by-diet interaction showed high genetic correlations for BW across diets, indicating stable genetic ranking and supporting the use of a common prediction model. In contrast, WGR and SGR exhibited evidence of diet-dependent genetic responses, suggesting the need for trait-specific modeling strategies. Genomic prediction analyses demonstrated that predictive performance varied across traits and diet groups, with BW achieving the highest accuracy. Notably, incorporating informative SNPs identified through GWAS markedly improved prediction accuracy, with values approaching 0.8. Evaluation of genotype-by-diet interaction showed high genetic correlations for BW across diets, indicating stable genetic ranking and supporting the use of a common prediction model. In contrast, WGR and SGR exhibited evidence of diet-dependent genetic responses, suggesting the need for trait-specific modeling strategies. Based on genomic estimated breeding values, superior broodstock were selected to establish an elite F1 population, enabling the practical application of genomic selection in breeding programs. Ongoing data collection from this population will further enhance model accuracy and robustness. Overall, this study demonstrates that genomic selection can effectively improve growth performance under varying feed conditions and provides a practical framework for developing feed-adaptive breeding strategies in olive flounder aquaculture.