Effects of Different Methods and Genomic Relationship Matrices on Reliabilities of Genomic Selection in Dairy Cattle
Livestock studies, vol.62, no.2, pp.58-64, 2022 (TRDizin)
- Publication Type: Article / Article
- Volume: 62 Issue: 2
- Publication Date: 2022
- Journal Name: Livestock studies
- Journal Indexes: TR DİZİN (ULAKBİM)
- Page Numbers: pp.58-64
- Open Archive Collection: AVESIS Open Access Collection
- Ankara University Affiliated: Yes
Abstract
Since genomic prediction is widely used in dairy cattle, we aimed to evaluate the performance of pedigree based (ABLUP), SNP based (GBLUP) and single-step GBLUP (ss-GBLUP) methods with different sets of information in terms of reliability of genomic prediction. Four different methods were evaluated: (Method 1) ABLUP with all available phenotypes and pedigree; (Method 2) GBLUP with SNP genotypes and phenotypes of genotyped cows; (Method 3) single-step GBLUP with SNP genotypes, phenotypes of genotyped cows and all pedigree and (Method 4) single-step GBLUP with SNP genotypes, all phenotypes of both genotyped and nongenotyped cows and all pedigree. SNP based methods also used different genomic relationship matrices (GRMs) formed by different approaches: vanRaden, Astle, Yang and Endelman. The simulated dataset replicates a common dairy cattle population. A significant increase in reliability of prediction was observed in ss-GBLUP with all phenotypes and pedigree beside genotyped cows. This increase was apparent for both first lactation milk yield (LMY) and milk fat percentage (Fat%). Combining all available information with ss-GBLUP gave about 1.6 and 1.2 times higher reliabilities for LMY and Fat%, respectively, compared to those obtained from the other three methods.