Genetic parameters and genomic prediction of egg production traits in ducks

Egg production traits are critical in duck breeding. Although genomic selection (GS) has been widely applied in livestock breeding, limited research has focused on predicting duck egg production traits, particularly across different physiological stages. In this study, we systematically explored the genetic parameters of egg production traits and evaluated the performance of GS in a commercial Pekin duck population. Analysis of 8,455 laying ducks showed that the heritability of cumulative egg production at 210, 280, and 300 days was 0.35, 0.37, and 0.48, respectively. Heritability during the early, peak, and post stages of cumulative egg production was 0.32, 0.15, and 0.24. The heritability of the egg laying rate from 25 to 60 weeks was 0.25. The cumulative egg production traits exhibited very strong mutual genetic correlations (≥ 0.95), while genetic correlations among egg laying stage traits ranged from 0.03 to 0.74. In the comparison between the pedigree-based best linear unbiased prediction (BLUP) and the genomic BLUP (GBLUP), using five-fold cross-validation, GBLUP outperformed the traditional pedigree BLUP model, with an average predictive reliability of 0.154, which was 0.029 higher than the predictive reliability of BLUP. In forward prediction, GBLUP also outperformed BLUP for all traits, with an average reliability of 0.097, which was 0.111 higher than the predictive reliability of BLUP. We also assessed the impact of linkage disequilibrium (LD) filtering on predictive reliability, which improved predictive reliability by 0.022 when the LD threshold was set to 0.14. In the comparison between GBLUP and Bayesian models utilizing a genotype with an LD pruning threshold of 0.14, GBLUP showed higher reliability than BayesB and BayesN in five-fold cross-validation, but lower than BayesCπ. In forward prediction, GBLUP demonstrated more robust performance, outperforming BayesB, BayesCπ, and BayesN, with improvements of 0.03, 0.019, and 0.05, respectively. This study provides a foundation for the application of GS in duck egg production and offers practical insights for improving predictive reliability in egg production.