Comparison of single-step methods for genomic prediction of age at first calving in dairy buffaloes

• Published in: The Journal of agricultural science 2024-11, p.1-7
• Main Authors: dos Santos, Jessica Cristina Gonçalves, de Araújo Neto, Francisco Ribeiro, Fernandez, Gabriela Stefani, de Abreu Santos, Daniel Jordan, Cunha, Felipe Pereira, Aspilcueta-Borquis, Rusbel Raul, Tonhati, Humberto
• Format: Article
• Language: English
• ISSN: 0021-8596; 1469-5146
• DOI: 10.1017/S0021859624000364

The age at first calving (AFC) is an important trait to be considered in breeding programmes of dairy buffaloes, where new approaches and technologies, such as genomic selection, are constantly applied. Thus, the objective of this study was to compare the predictive ability of different genomic single-step methods using AFC information from Murrah buffaloes. From a pedigree file containing 3320 buffaloes, 2247 cows had AFC records and 553 animals were genotyped. The following models were performed: pedigree-based BLUP (PBLUP), single-step GBLUP (ssGBLUP), weighted single-step GBLUP (WssGBLUP), and single-step Bayesian regression methods (ssBR-BayesA, BayesBπ, BayesCπ, Bayes-Lasso, and BayesRR). To compare the methodologies, the accuracy and dispersion of (G)EBVs were assessed using the LR method. Accuracy estimates for the genotyped animals ranged from 0.30 (PBLUP) to 0.39 (WssGBLUP). Predictions with the traditional model (PBLUP) were very dispersed from what was expected, while BayesCπ (0.99) and WssGBLUP (1.00) obtained the lowest dispersion. The results indicate that the use of genomic information can improve the genetic gain for AFC by increasing the accuracy and reducing inflation/deflation of predictions compared to the traditional pedigree-based model. In addition, among all genomic single-step models studied, WssGBLUP and single-step BayesA were the most advantageous methods to be used in the genomic evaluation of AFC of buffaloes from this population.