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 #392062

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: Flexible testing and use of milk-only records

item Vanraden, Paul
item Fok, Gary
item BACHELLER, LILLIAN - Council On Dairy Cattle Breeding
item JANSEN, GERALD - Council On Dairy Cattle Breeding
item CARRILLO, JOSE - Council On Dairy Cattle Breeding

Submitted to: Journal of Dairy Science
Publication Type: Abstract Only
Publication Acceptance Date: 3/10/2022
Publication Date: 6/19/2022
Citation: Van Raden, P.M., Fok, G.C., Bacheller, L.R., Jansen, G.B., Carrillo, J.A. 2022. Flexible testing and use of milk-only records [abstract]. Journal of Dairy Science. 105(Suppl. 1):199(abstr. 1498V).

Interpretive Summary:

Technical Abstract: National genetic evaluation software assumed that fat yield was always recorded and excluded milk-only records. Milking systems often can accurately measure and record milk volume, but inline estimation of milk components is more difficult. Edit programs were revised to begin using milk-only records in April 2022 in the multi-trait evaluation of yield traits. Other trait groups such as cow fertility and health also will have more usable records because edits for those traits require having usable yield records. Herd variance ratios had been estimated from 1 trait (milk yield since 1992 and then fat yield since 2007) and applied to adjust all 3 traits (milk, fat, and protein). Since 1998 when data collection ratings (DCR) were introduced, milk and fat received the same weight calculated as an average of the DCR values for milk and components instead of separate weights because of software limitations. Programs were revised to use trait-specific variance adjustments and weights, to include milk-only records, and to remove much obsolete code. Most milk-only records are unsupervised and therefore get the same reduced weights and extra edits for percent milk shipped and percentage of valid ID as other owner-sampler herds. Lactation weights for milk, fat, and protein now use 3 separate DCR based on the testing patterns and correlations among test days within lactations. New and official genetic evaluations were compared from December 2020 data. Numbers of usable lactation records were 98,269,605 for milk, 97,393,419 for fat, and 78,044,073 for protein, indicating that 876,186 milk-only records were added. Correlations of new with previous PTA were > 0.9995 across all bulls for all 3 traits and were > 0.997 for bulls born since 2007 with > 50% reliability. The SD of PTA increased slightly by 1.3% for milk, 2.1% for fat, and 2.4% for protein but reliability also increased a little from the extra records. Further research could help adapt to more flexible testing options and automated data collection that continue to increase in popularity.