Genetic and genomic evaluations of quantitative milking speed phenotypes

Authors: Miles, Asha; Hutchison, Jana; Toghiani, Sajjad; O’CONNELL, JEFF - University Of Maryland School Of Medicine; FOURDRAINE, ROBERT - Dairy Records Management Systems(DRMS); Van Raden, Paul; PARKER GADDIS, KRISTEN - Council On Dairy Cattle Breeding; SIEVERT, STEVEN - Collaborator; EAGLEN, SOPHIE - National Association Of Animal Breeders; BEWLEY, JEFFREY - Holstein Association Usa, Inc; DURR, JOAO - Council On Dairy Cattle Breeding

Submitted to: International Committee on Animal Recording(ICAR)
Publication Type: Proceedings
Publication Acceptance Date: 9/25/2024
Publication Date: 1/31/2025
Citation: Miles, A.M., Hutchison, J.L., Toghiani, S., O’Connell, J.R., Fourdraine, R.H., Van Raden, P.M., Parker Gaddis, K.L., Sievert, S., Eaglen, S., Bewley, J., Durr, J.W. 2025. Genetic and genomic evaluations of quantitative milking speed phenotypes. International Committee on Animal Recording (ICAR). ICAR Technical Series no. 28, p. 359-362.

Interpretive Summary:

Technical Abstract: Many milking systems with inline milk meters can record the milk yield and duration of each milking for individual cows. The objective of this work was to determine the suitability of milking speed traits for genetic and genomic selection and the amount of phenotype data required to produce a reliable evaluation. Records from January 2021 to December 2022 were retrieved by Dairy Records Management Systems, comprising data from 305 herds, 9 different original equipment manufacturers and 23,201 complete lactations of 23,180 cows, including 4,246 genotyped cows. Milking speed was defined as milk yield divided by milking duration for each individual milking. Four traits were compared: 1) average of total lactation data for all parities, 2) average of test days for all parities, 3) average of total lactation data for first parity only, 4) average of test days for parity 1. Breed, milking frequency, parity, lactation length, and meter manufacturer were included in the genetic model along with genetic groups and permanent environment. The pedigree relationship matrix included 219,703 animals with records or descendants with records plus 96 million other animals. Variances were estimated by both Gibbs sampling and REML; estimates were very similar. Residual variance was 51% higher for test day traits compared to total lactation traits. Milking speed test day heritability was 28% vs. 37% for total lactation data; genetic correlation between them was 0.97, suggesting that even with a 99% reduction in amount of phenotypic data included they are describing the same trait. Milking speed was less stable in parity 1 compared to other parities, but high genetic correlations (> 0.92) suggest the same trait is being captured. Milking speed had a small favorable genetic correlation with milk yield but unfavorable with somatic cell score based on 756 Holstein bulls with reliability > 50%. Genomic predictions for young animals born in the last 10 years averaged 37% reliability compared to ~70% reliability for several other traits. We conclude that evaluations for milking speed are not only feasible but would have significant economic impact for producers using various milking systems. Work on implementing an evaluation for milking speed is currently underway.