Evaluation of prediction accuracy of the genomic breeding values of some economic traits in Iranian Holstein cattle using SNP markers and haplotype blocks

The aim of current study was to evaluate the accuracy of genomic breeding values (GEBV) for two important economical traits of milk yield and somatic cell score using SNP markers and LD-based haplotype blocks (haploblocks) by two statistical methods of GBULP and Bayes B. The data set consisted of 1654 bulls genotyped with different marker densities. When SNPs were used, the accuracy of breeding values obtained by Bayes B was better than GBLUP. In other words, for milk yield and somatic cell score traits, the prediction accuracy of GBLUP was 0.54 and 0.44 and by Bayes B was 0.58 and 0.44,respectively. For milk yield, the prediction accuracy of using haploblocks in both statistical methods was higher than the prediction accuracy using SNPs, while for the somatic cell score, this increase was more pronounced when GBLUP was used. However, when Bayes B was used this superiority was only obtained when the r2 statistic used to build the haploblocks was higher than 0.2. The results showed that the optimum level of r2 for building haploblocks depends on the trait type and its heritability. As a result, using r2 statistic more than 0.2 for building haploblocks can increase the accuracy of breeding valuesfoe both traits compared to SNP markers.