Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Animal Genomics and Improvement Laboratory » Research » Publications at this Location » Publication #392060
Research Project: Improving Dairy Animals by Increasing Accuracy of Genomic Prediction, Evaluating New Traits, and Redefining Selection Goals

Location: Animal Genomics and Improvement Laboratory

Title: Accuracy of genomic predictions including or excluding foreign data in reference populations

Author
item MOTA, RODRIGO - Council On Dairy Cattle Breeding
item Vanraden, Paul

Submitted to: Journal of Dairy Science
Publication Type: Abstract Only
Publication Acceptance Date: 3/10/2022
Publication Date: 6/19/2022
Citation: Mota, R.R., Van Raden, P.M. 2022. Accuracy of genomic predictions including or excluding foreign data in reference populations [abstract]. Journal of Dairy Science. 105(Suppl. 1):130(abstr. 1330).

Interpretive Summary:

Technical Abstract: Genomic predictions often include foreign data obtained from Interbull’s multi-trait across-country evaluation (MACE), but accuracy advantages may be reduced as domestic datasets grow and many cows are genotyped. To test this, accuracies of predictions were compared from truncated domestic data vs. a combined reference that also included truncated MACE (TMACE) data. The predictions used the current multi-step models instead of testing with previously official methods that may have been revised during the past 4 years. Predictions were tested for 10 traits and Net Merit $. Genomic estimated breeding values (GEBV) were compared for 7,651 bulls born 2012 or later that had no daughter records in August 2017 but had >75% reliability of conventional EBV in August 2021. The validation bulls consisted of 4 groups: 3,623 domestic Holsteins and 544 Jerseys with USA (or country code 840) identification and 3,386 foreign Holsteins and 98 Jerseys with other country identification. Smaller subsets of these bulls were used for traits with less reliability. Most squared correlations (R2) were unexpectedly higher for predicting foreign than USA bulls, possibly because the foreign bulls were less selected or obtained less reliability than the USA bulls on USA scale. Inclusion of TMACE gave largest benefits to predict less heritable traits such as productive life and somatic cell score in both breeds. Gains were larger for foreign bulls than for the same traits of USA bulls. For predicting higher heritability yield traits, benefits were near 0 for USA bulls, small (1 to 2%) for foreign Holsteins, and larger (5 to 11%) for foreign Jerseys. All regressions (B1) were near 1.0 and changed only a little with the inclusion of TMACE. MACE can increase prediction reliability even when large domestic reference populations are available, and results from TMACE more accurately demonstrate that by comparing predictions using the same current model.