IMPROVING GENETIC PREDICTIONS FOR DAIRY ANIMALS USING PHENOTYPIC AND GENOMIC INFORMATION
Title: Multibreed genomic evaluations using purebred Holsteins, Jerseys, and Brown Swiss
Submitted to: Journal of Dairy Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: March 19, 2012
Publication Date: September 1, 2012
Citation: Olson, K.M., Van Raden, P.M., Tooker, M.E. 2012. Multibreed genomic evaluations using purebred Holsteins, Jerseys, and Brown Swiss. Journal of Dairy Science. 95(9):5378-5383.
Interpretive Summary: Genotypes of purebred Holstein, Jersey, and Brown Swiss were used to predict of genomic merit across all breeds. Three different evaluation methods were investigated. A method that used only the prediction equation from another breed worked poorly. A method that combined all breeds into one population had higher accuracies than parent average but was not as accurate as results within breed. A correlated trait model where traits were the marker effects in each breed increased the accuracy of genomic predictions above the within breed estimates and outperformed the other methods used.
Multibreed models are currently used in traditional USDA dairy cattle genetic evaluations of yield and health traits, but within-breed models are used in genomic evaluations. Multibreed genomic models were developed and tested using the 19,686 genotyped bulls included in the official August 2009 USDA genomic evaluation. The data were divided into training and validation sets. The training data set was comprised of bulls that were proven (had daughter information) as of November 2004 and totaled 5,331, 1,361, and 506 Holstein, Jersey, and Brown Swiss, respectively. The validation data set had 2,508 Holstein, 413 Jersey, and 185 Brown Swiss bulls that were unproven (no daughter information) in November 2004 and proven by August 2009. A common set of 43,385 single nucleotide polymorphisms (SNP) was used for all breeds. Three methods of multibreed evaluation were investigated. Method 1 estimated SNP effects separately within-breed and then applied those breed specific SNP estimates to the other breeds. Method 2 estimated a common set of SNP effects from combined genotypes and phenotypes of all breeds. Method 3 solved for correlated SNP effects within each breed estimated jointly using a multitrait model where breeds were treated as different traits. Regressions were used to test across-breed genomic predicted transmitting ability (GPTA) with the traditional within-breed GPTA. Method 1 worked poorly and correlations between using within breed and using other breeds’ SNP estimates were less than 30%. Coefficient of determinations (R2) for within breed predictions were higher than using another breeds’ SNP estimates. Across-breed GPTA from method 2 had higher R2 values than parent average alone but typically produced lower R2 values than the within breed GPTA. A few traits in the Brown Swiss breed had a higher accuracy with the across-breed GPTA, probably due to small numbers of animals. Correlations between within-breed GPTA and across-breed GPTA ranged between 0.91 and 0.93. Results for the multi-breed GPTA from method 3 were significant, and adjusted coefficient of determinations for protein yield (the only trait tested for method 3) were highest of all methods for all breeds. However, compared with the current within-breed genomic model, method 3 increased the adjusted coefficient of determination by only 0.010, 0.004, and 0.002 for Brown Swiss, Jerseys, and Holsteins, respectively.