Abstract:
The accuracy of breeding value affects to the response to selection. Due to the estimation of breeding value (EBY) from pedigree took long period to collect the selection data, thus the genomic data were used to reduce duration of data collection and increase the accuracy of selection. The important factors which influents to the accuracy of genomic breeding value (GBV) were size of reference population, number of SNPs and heritability of the trait. The objectives this study were to analyze relationshup between reference population size and accuracy of genomic breeding values of dairy cattle population in Upper Northem Thailand. The data was simulated the reference population or calibration set ( CS), which the population were known phenotype. Furthermore, validation set (VS) was unknown phenotype population. Sumulation used information of dairy cattle population, during 1952 to 2015. Accuracy amalysis was used correlation between EBY, GBV and TBV, applying with animal model. The 0.1, 0.3 and 0.5 of heritabity and number of SNPs 3K, 7K, 15K, 30K, 50K and 60K were used to simulation. The results showed that the accuracy of GBV higher than EBV and the calibration set accuracy higher than validation set. In addion, the accuracy of low heritability trait is not remaining stable. The 100-1,500 of reference population size were estimated accuracy with SNPs at heritability of 0.3 and 0.5. The highest acy of calibration set estimated from pedigree data were 0.721 and 0.841 in 300 and 200 animals respectively. While, the highest accuracy estimated from genormic data were 0.751 and 0.855 in 1,500 and 1.500 animals respectively. However, the highest accuracy of validation set estimated from pedigree data were 0.483 and 0.519 in 300 and 200 animals respectively. Whereas, the highest accuracy estimated from genomic data were 0.554 and 0.679 in 1,400 and 1,500 animals respectively. Therefore, genomic selection was high accuracy of selection that affected to the genetic progress compared with conventional breeding.