PS1-3. Comparison of Genomic Predictions from Single- and Multi-trait GBLUP Including Additive and Dominance Effects for of Body Weights in Brangus Population.

Abstract The use of multi-trait approaches to genetic evaluation has been advocated not only because it is more accurate and precise, but also because even moderately inheritable traits can benefit from information from correlated traits. Traditional multi-trait genetic evaluation models (GBLUP) used to estimate genetic parameters include only additive genetic effects; however, many studies have shown that dominance genetic effects are an essential component of genetic evaluation. This study aimed to compare the predictive ability of single- and multi-trait GBLUP models including additive and dominance effects for the following body weights: birth weight (BirthWt), weaning weight (WeanWt) and yearling weight (YearWt) from 740 Brangus heifers that were registered with International Brangus Breeders Association. After quality control of SNP markers generated with Illumina vl 50K BeadChips, 35351 SNP markers coded as 0, 1, or 2, based on the number of B alleles at each locus were used in the single- and multi-trait GBLUP models to obtain additive and dominance heritabilities and genetic correlation estimates between BirthWt, WeanWt and YearWt traits. The package BGLR (definition?) with Multitrait function in R language was used to obtain heritability estimates and genetic correlations between BirthWt, WeanWt and YearWt traits. Brangus heifers were randomly clustered into 10-fold cross-validation to quantify the correlations between actual and predicted phenotypic values from single- or multi-trait GBLUP models including additive and dominance effects. Additive and dominance heritability estimates from multi-trait GBLUP were 0.20, 0.14 and 0.21; and 0.15, 0.18 and 0.19, whereas the single-trait GBLUP estimates were 0.21, 0.14 and 0.22; and 0.14, 0.17 and 0.20 for BirthWt, WeanWt and YearWt traits, respectively. Additive and dominance heritability estimates from single- and multi-trait GBLUP models were quite similar for BirthWt, WeanWt and YearWt traits. In the single- and multi-trait GBLUP analysis, additive heritability estimates were higher than dominance heritability estimates for BirthWt, lower than that of WeanWt and similar to that of YearWt. Genetic correlations between BirthWt and WeanWt (0.34), and between BirthWt, YearWt (0.21) were low; however, genetic correlation between WeanWt and YearWt (0.67) was moderate. Genetic correlations indicated the positive genetic association between BirthWt, WeanWt and YearWt traits. The prediction abilities of single-trait GBLUP with dominance effect (0.201, 0.112 and 0.256) and multi-trait GBLUP with dominance effect (0.191, 0.112 and 0.256) in cross-validation data sets were quite similar for BirthWt, WeanWt and YearWt traits, respectively.