Abstract: The purpose of this study, was to determine the accuracy of imputation for non-genotyped sires using their progenies genotypes and to compare the accuracy of genomic breeding values from imputed sires with respect to their true genomic data. A total of 1,800 were simulated to construct phenotypic data and pedigree data. A genotype panel for all animals was prepared as well. Among them, 20 sires were selected randomly and imputed assuming that their genotypes were missing. The average accuracy of imputation for sires through 100 simulation repeats was 88.7% and the obtained range of accuracies was 87-90.4%. The accuracy of Genomic Breeding Values (GEBV) of whole population was slightly higher than for pedigree based Breeding Values (EBV) 0.639 and 0.611, respectively. The accuracy for GEBV from 20 selected sires using true or imputed genotypes were 0.673 and 0.669, respectively. These results indicated that imputation of missing sire genotypes can obtain an accuracy which is almost close to the accuracy of estimates from their true genotypes. Therefore, it can be concluded that there is a possibility in genomic selection for using imputed sires which have no genotypes but yet have their progenies genotypes available. These results deemed more applicable in Korean dairy genomic evaluation, i.e., for Holstein where most sire genotypes are difficult to obtain due to an extensive use of imported semen.
C.I. Cho, M. Alam, T.J. Choi and K.H. Cho, 2013. Accuracy of Genomic Selection Through Imputation of Sires in Dairy Cattle. Research Journal of Dairy Sciences, 7: 14-17.