Location: Livestock and Range Research LaboratoryTitle: Multi-trait predictions using GBLUP and Bayesian mixture prior model in beef cattle
|WANG, ZEZHAO - Chinese Academy Of Agricultural Sciences|
|MA, HAORAN - Chinese Academy Of Agricultural Sciences|
|LI, HONGWEI - University Of Alberta|
|XU, LEI - Anhui Academy Of Agricultural Sciences|
|ZHU, BO - Chinese Academy Of Agricultural Sciences|
|Hay, El Hamidi|
|XU, LINGYANG - Chinese Academy Of Agricultural Sciences|
|LI, JUNYA - Chinese Academy Of Agricultural Sciences|
|LI, HONGYAN - Tongliao Animal Agriculture Development Service Center|
Submitted to: Animal Research and One Health
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/29/2023
Publication Date: 8/6/2023
Citation: Wang, Z., Ma, H., Li, H., Xu, L., Zhu, B., Hay, E.A., Xu, L., Li, J., Li, H. 2023. Multi-trait predictions using GBLUP and Bayesian mixture prior model in beef cattle. Animal Research and One Health. https://doi.org/10.1002/aro2.13.
Interpretive Summary: Genetic improvement of complex traits such as fertility has been hindered by the low heritability of these traits. Therefore, the objective of this study is to evaluate the ability of a multiple-trait approach using genomic information to improve prediction accuracy of the genetic merit of several traits with varying heritability levels and increase the overall genetic gain. Our study suggests that multi-trait prediction using conventional models and our proposed model is feasible for genomic selection in beef cattle. Our findings also indicate that our method outperforms other existing models especially for the low-heritability traits and shows high reliability in prediction accuracy.
Technical Abstract: Background: The approach of multiple-trait genomic selection (MTGS), which incorporates correlated traits can improve the prediction ability of low heritability traits. Previous MTGS methods based on BLUP assume that a single SNP marker simultaneously affects multiple traits. However, this assumption ignores the real genetic architecture of multiple traits. Therefore, Bayesian multiple-trait methods incorporating mixture priors have been proposed to account for the true genetic architecture and their relationship between genotypes and traits of interest. In this study, we evaluated genomic prediction accuracy using multi-trait BayesCp method (MT-BayesCp), which allows for a broader range of mixture priors for economically important traits in beef cattle. We then compared the prediction performance of MT-BayesCp with single-trait GBLUP (ST-GBLUP), multi-trait GBLUP (MT-GBLUP) and single-trait BayesCp (ST-BayesCp) methods. Results: In this study, we found that ribeye area (REA) and ribeye weight (REWT) showed high heritability, while slaughter weight (SWT) and carcass weight (CWT) displayed moderate heritability, and slaughter rate (SR) and feedlot average daily gain (FDG) showed low heritability using all four approaches. High positive genetic correlations (>0.90) were observed between CWT and SWT (0.981), and SR and REWT (0.921). Notably, the MT-BayesCp method showed superior predictive abilities compared to other models in terms of prediction accuracy and prediction bias. Using the MT-BayesCp method, the accuracy increased from 0.272 to 0.694 for CWT compared to ST-GBLUP and ranged from 0.632 to 0.694 compared to ST- BayesCp. MT-GBLUP and ST-BayesCp showed similar prediction accuracies, while MT- BayesCp showed the least biased evaluations. Additionally, our results suggested that prediction accuracy is improved with the increase of heritability of traits. Prediction accuracy of low-heritability traits significantly increased when they were combined with correlated traits with high genetic correlation in a multi-trait prediction approach. Conclusions: Our study suggests that multi-trait predictions using conventional BLUP and Bayesian mixture prior models is feasible for genomic selection in beef cattle. Our findings also indicate that the MT-BayesCp method outperforms other models (ST-GBLUP, MT-GBLUP and ST- BayesCp), especially for the low-heritability traits and shows high reliability in prediction accuracy.