Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Animal Genomics and Improvement Laboratory » Research » Publications at this Location » Publication #293635

Title: Accuracy of genomic predictions in Bos indicus (Nellore) cattle

Author
item NEVES, HAROLDO - Universidade Estadual Paulista (UNESP)
item CARVALHEIRO, ROBERTO - Universidade Estadual Paulista (UNESP)
item PEREZ O'BRIEN, ANA - University Of Natural Resources & Applied Life Sciences - Austria
item UTSUNOMIYA, YURI - Universidade Estadual Paulista (UNESP)
item DO CARMO, ADRIANA - Universidade Estadual Paulista (UNESP)
item SCHENKEL, FLAVIO - University Of Guelph
item SOLKNER, JOHANN - University Of Natural Resources & Applied Life Sciences - Austria
item NCEWAB, JOHN - Agresearch
item Van Tassell, Curtis - Curt
item Sonstegard, Tad
item Cole, John
item DA SILVA, MARCOS - Embrapa
item QUEIROZ, SANDRA - Universidade Estadual Paulista (UNESP)
item GARCIA, JOSE - Universidade Estadual Paulista (UNESP)

Submitted to: Genetic Selection Evolution
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/18/2014
Publication Date: 2/27/2014
Publication URL: https://handle.nal.usda.gov/10113/58866
Citation: Neves, H.H., Carvalheiro, R., Perez O'Brien, A.M., Utsunomiya, Y.T., Do Carmo, A.S., Schenkel, F.S., Solkner, J., Ncewab, J.C., Van Tassell, C.P., Sonstegard, T.S., Cole, J.B., Da Silva, M.V., Queiroz, S.A., Garcia, J.F. 2014. Accuracy of genomic predictions in Bos indicus (Nellore) cattle. Genetic Selection Evolution. 46:17.

Interpretive Summary: The Nellore is a tropically adapted breed of cattle that has an important role in beef production in tropical systems, such as in Brazil. There is considerable interest from Nellore producers in the use of genomic evaluation to increase rates of genetic gain, but the lack of bulls with traditional estimated breeding values with high accuracies presents a substantial challenge to implementation of such a program. This study describes initial results for a genomic selection program based on 685 bulls with high-density single nucleotide polymorphism genotypes. Phenotypes included body weight and carcass traits, gestation length, scrotal circumference, temperament, and two selection indices. A forward-prediction scheme was adopted, in which sires with highly accurate proofs in 2007 composed the training group, and sires without proofs in 2007 but with accurate proofs in 2011 composed the testing group. Accuracies of genomic predictions ranged from 0.14 (temperament) to 0.70 (finishing precocity evaluated through visual scores), and were very consistent across prediction methods. These results demonstrate the technical feasibility of genomic selection in a Nellore population, although further research needs to be conducted to enable more cost-effective selection decisions using genomic information.

Technical Abstract: Nellore cattle play an important role in beef production in tropical systems, and there is great interest in determining if genomic selection can contribute to its improvement. The first results of the application of genomic selection in a Bos indicus (Nellore) population are presented in this paper. Influential bulls were genotyped with the Illumina Bovine HD assay in order to assess genomic predictive ability for body weight and carcass traits, gestation length, scrotal circumference, temperament, and two selection indices. After quality controls, 685 samples and 320,238 SNP remained in the analyses. A forward-prediction scheme was adopted, where sires with highly accurate proofs in 2007 composed the training group and sires without proofs in 2007 but with accurate proofs in 2011 composed the testing group. In the training step, bulls’ estimated breeding values (EBV) were used as pseudo-phenotypes for estimation of marker effects using four different methods: GBLUP, ridge regression, mixture model (Bayes C), and Bayesian LASSO. Accuracies of genomic predictions were assessed through the correlation between direct genomic breeding values (DGV) and 2011 EBV for the testing group. Accuracies of genomic predictions ranged from 0.14 (temperament) to 0.70 (finishing precocity evaluated through visual scores) and were very consistent across prediction methods. Averages of individual accuracies of DGV, estimated using GBLUP, were consistent across traits and averaged around 0.46. Further analyses suggested that higher-than-expected accuracies were observed for traits affected by genotype stratification. Two subgroups of the sampled population were examined with a principal components analysis based on genomic kinship coefficient. Genomic BLUP predictions slightly outperformed the predictions obtained with two variable selection methods (Bayesian LASSO and Bayesian mixture model). The accuracies of genomic predictions in this population are dependent on the degree of relatedness between animals in the training and testing sets. The technical feasibility of genomic selection application in a Bos indicus (Nellore) population was demonstrated, although further research on incorporating this technology in breeding schemes needs to be conducted to enable more cost-effective selection decisions using genomic information.