Study of genomic prediction methods-Comparisons of traditional BLUP, G-BLUP and the single-step BLUP

The purpose of this study was to compare the accuracy of three genomic prediction methods based on traditional BLUP, G-BLUP and single step method ( or SS-BLUP) for milk (MY), fat (FY), protein yield (PY) and somatic cell count (SCC) in Iranian Holstein dairy cattle. In this study the genomic data records of 345 Iranian Holstein dairy cattle, genotyped using the Illumina Bovine SNP50 BeadChip, were used. Genomic breeding values for the studied traits was estimated by MIXBLUP software. Results showed that the mean of genomic prediction accuracy of traditional BLUP, G-BLUP and SS-BLUP was 0.39, 0.47 and 0.54 respectively. Accuracy estimation ranged from 0.32 for somatic cell count trait in traditional BLUP method to 0.59 for milk production trait in SS-BLUP method. Genomic prediction accuracies under SS_BLUP for the considered traits including MY, FY, PY and SCC were 0.1, 0.07, 0.09 and 0.07 higher than the corresponding values obtained under G-BLUP, respectively. The corresponding values under SS-BLUP were 0.16, 0.14, 0.17 and 0.14 higher than the values obtained under traditional BLUP. Mean squared error for SS-BLUP compared to other methods was lower for all traits. Genomic prediction regression coefficients under tradition BLUP, G-BLUP and SS-BLUP methods were 0.77, 0.88 and 0.94, respectively. Therefore, compared to traditional BLUP and G-BLUP, the SS-BLUP approach can significantly improve the accuracy of genomic prediction for milk production traits in Iranian Holstein cattle.