Genomic predictions for growth and feed efficiency traits in a duck breeding population.

Abstract

Background: In the commercial broiler duck industry, optimizing breeding practices is crucial, especially for growth and feed efficiency traits. Although genomic selection (GS) has been successfully applied in livestock, its use in duck breeding is not yet widespread. This study aims to investigate genetic parameters and refine GS strategies for feed efficiency and growth traits in ducks, paving the way for more precise and efficient breeding programs.

Results: We investigated genetic parameters of 12 growth and feed efficiency traits in a commercial breeding Line of 52,610 ducks across 10 generations. We applied genomic predictions to 2779 ducks from the latest three generations. Heritability of these traits ranging from 0.16 to 0.51. Genomic prediction accuracy was higher for GBLUP under cross-validation than forward validation. This performance discrepancy was influenced by reference population recency and trait complexity. Notably, single-step GBLUP consistently outperformed pedigree-based BLUP, particularly for feed efficiency traits. Expanding the reference population with recent generations improved forward validation accuracy by 27.7%, highlighting the critical role of updated genetic data in enhancing across-generation predictive accuracy. The newly proposed residual feed intake adjusted for breast muscle volume demonstrated a higher heritability and predictive accuracy compared to its predecessor. Pruning variants using Linkage disequilibrium thresholds of 0.075 resulted in an increase of 0.05 in the average predictive accuracy. Similarly, omitting the Hardy-Weinberg equilibrium threshold generally resulted in higher predictive accuracy for most traits. However, for traits such as BMW, BMT, and BMV, we observed enhanced predictive accuracy when applying a specific threshold for HWE test pruning. The BayesRC model, when informed by cis-eQTLs or their regulated genes, particularly from adipose and muscle tissues, increased predictive accuracy for various traits, highlighting the importance of integrating biological data into genomic prediction frameworks.

Conclusions: This study offers encouraging evidence for utilizing GS to enhance growth and feed efficiency traits in ducks. It offers valuable insights into optimizing GS for duck breeding, emphasizing the critical roles of model selection, marker density refinement, and the strategic integration of prior biological information.