Submitted to: Journal of Dairy Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: July 25, 2011
Publication Date: December 1, 2011
Citation: Wiggans, G.R., Cooper, T.A., Van Raden, P.M., Cole, J.B. 2011. Technical note: Adjustment of traditional cow evaluations to improve accuracy of genomic predictions. Journal of Dairy Science. 94(12):6188-6193.
Interpretive Summary: A method of adjusting cow genetic evaluations was developed that increased the accuracy of the genomic evaluations. Cow evaluations were adjusted to have similar characteristics to those of bulls. Cows were grouped on deviation of their evaluation from direct genomic value calculated using only bulls as predictors. Within these groups, mean and variance adjustments were calculated as a function of daughter equivalents. This system allowed for adjustment to vary by cow, an improvement over the previous adjustment where all cows received the same adjustment. Genomic evaluations are more accurate by better using the information from cows.
Genomic evaluations are calculated using values that have been deregressed from traditional PTA to estimate single nucleotide polymorphism (SNP) effects. Previous research indicated that including cow genomic data to calculate SNP effects did not increase reliabilities of genomic evaluations of yield traits. Upward bias in traditional PTA of genotyped cows was proposed as the cause. The direct genomic value (DGV) is the sum of an animal’s SNP effects, and should be consistent with traditional PTA. This is the case for bulls, but the traditional PTA for genotyped cows are higher than those of the bulls. To make the cow PTA more like those of the bulls for the yield traits (milk, fat and protein), mean and variance adjustments were calculated. Cow Mendelian sampling (PTA – PA) was adjusted to be similar to bulls. Two methods of adjustment were investigated. Method 1 used the same mean and variance adjustment for all cows in a breed. Method 2 assigned cows to one of five groups based on the difference between their PTA and DGV based only on bull evaluations. It also varied the adjustment by the daughter equivalents of the cow’s evaluation with the contribution from parents removed. The objective of method 2 was to tailor the adjustment to the individual cow to better accommodate the large number of commercial cows expected to be genotyped with the less expensive 3K SNP chip. To determine gains in reliabilities, predictions were made for bulls with current evaluations that did not have evaluations in August 2006. The predicted values were compared with the bull’s actual evaluation from August 2010. Genomic evaluations for Holsteins, Jerseys, and Brown Swiss can be more accurate by better using the information from cows.