Summary and Implications
The objective of this study was to estimate accuracies of genomic predictions based on 50K SNP genotypes for 8 nationally evaluated traits in US Maine-Anjou beef cattle using single or multi-breed reference populations. The accuracies of direct genomic values (DGV) ranged from 0.22 to 0.45 for 8 studied traits when the reference populations comprised only 573 Maine-Anjou animals. Accuracies were slightly reduced and ranged from 0.21 to 0.38 when the reference population included over 9,000 animals from many other breeds as well as Maine-Anjou. These results demonstrate that including data from other populations does not generally increase accuracy of prediction in one particular population. This means every breed association must develop its own reference population if it intends to offer genomic prediction.
Iowa State University
Lee, Jungjae; Saatchi, Mahdi; Su, Hailin; Fernando, Rohan L.; and Garrick, Dorian J.
"Genomic Prediction using Single or Multi-Breed Reference Populations in US Maine-Anjou Beef Cattle,"
Animal Industry Report:
AS 660, ASL R2856.
Available at: http://lib.dr.iastate.edu/ans_air/vol660/iss1/21