Accuracy of genomic prediction for milk production traits with different approaches in a small population of Slovenian Brown bulls

详细信息    查看全文

• 作者：M. 艩pehar ; K. Poto膷nik ; G. Gorjanc
• 关键词：Genomic selection ; Breeding value ; Accuracy ; Bivariate model
• 刊名：Livestock Science
• 出版年：November, 2013
• 年：2013
• 卷：157
• 期：2-3
• 页码：421-426
• 全文大小：662 K

文摘

Use of genome-wide single-nucleotide polymorphism (SNP) data enables accurate estimates of breeding value (EBV) for young animals when a sufficiently large number of animals in a population are genotyped. The objective of this study was to evaluate the accuracy of integrating genomic data into the national evaluation system for milk traits in a small population of Slovenian Brown bulls using univariate national evaluation based on phenotype and pedigree data (U), international direct genomic value (DGV), and bivariate national evaluation incorporating DGV as a correlated trait (B). Comparison of approaches was assessed separately for training and validation subset of bulls using theoretical and empirical accuracy. Genetic correlation between the phenotype based EBV and DGV was between 0.79 and 0.86 confirming the utility of DGV for prediction. Use of DGV did not improve already high accuracy (0.98) for proven bulls due to the substantial number of daughters per bull. In young bulls, inclusion of DGV in B analysis has increased theoretical accuracy of prediction from 0.58 to 0.89 and further from 0.92 to 0.96 when these bulls were progeny tested. Empirical correlations on the validation subset confirmed the observed increases in theoretical accuracy although values were considerably lower due to the low and variable number of daughters per bull in the validation subset. When combining both progeny and DGV data, correlation between the U and B evaluation was 0.92 in validation subset confirming the usefulness of integrating both data sources. Integration of all the available information is not only beneficial for the use of all the data, but also to simplify publication since all information can be combined in a single breeding value.