Accuracy of genomic breeding values and predictive ability for postweaning liveweight and age at first calving in a Nellore cattle population with missing sire information

Abstract

The multiple sire system (MSS) is a common mating scheme in extensive beef production systems. However, MSS does not allow paternity identification and lead to inaccurate genetic predictions. The objective of this study was to investigate the implementation of single-step genomic BLUP (ssGBLUP) in different scenarios of uncertain paternity in the evaluation for 450-day adjusted liveweight (W450) and age at first calving (AFC) in a Nellore cattle population. To estimate the variance components using BLUP and ssGBLUP, the relationship matrix (A) with different proportions of animals with missing sires (MS) (scenarios 0, 25, 50, 75, and 100% of MS) was created. The genotyped animals with MS were randomly chosen, and ten replicates were performed for each scenario and trait. Five groups of animals were evaluated in each scenario: PHE, all animals with phenotypic records in the population; SIR, proven sires; GEN, genotyped animals; YNG, young animals without phenotypes and progeny; and YNGEN, young genotyped animals. The additive genetic variance decreased for both traits as the proportion of MS increased in the population when using the regular REML. When using the ssGBLUP, accuracies ranged from 0.13 to 0.47 for W450 and from 0.10 to 0.25 for AFC. For both traits, the prediction ability of the direct genomic value (DGV) decreased as the percentage of MS increased. These results emphasize that indirect prediction via DGV of young animals is more accurate when the SNP effects are derived from ssGBLUP with a reference population with known sires. The ssGBLUP could be applied in situations of uncertain paternity, especially when selecting young animals. This methodology is shown to be accurate, mainly in scenarios with a high percentage of MS.