Genetic analysis of daily milk weights in US Holsteins using pen-based contemporary groups

Abstract

The availability of daily milk weights and pen location information provides an interesting opportunity to capture additional data and review how contemporary groups are defined for dairy cattle genetic evaluations. In the United States, dairy cows in larger herds are grouped into pens according to various characteristics such as parity, production level, reproductive status, lactation stage, and health status. Our dataset included pen location information for each daily milk weight, so instead of using herd-year-season of calving to form contemporary groups, we used herd-pen-milking date to more precisely model the environmental effects cows experience at the pen level on a given day. Our dataset included 21,000,951 aggregated daily milk records from 114,243 first-parity Holstein cows milked 3 times daily in conventional parlor systems in 157 herds representing 29 US states. Our phenotype of interest was daily milk weight, and alternative repeatability animal models were used to estimate genetic parameters and predict breeding values. Age at first calving (6 levels) and DIM (10 levels) were included as fixed effects and cow (114,243 levels) was included as a random effect. Contemporary group effects included a fixed or random herd-year-season of calving effect (1,492 levels) and a fixed or random herd-pen-milking date effect (285,592 levels). Genetic parameters (kg²; posterior SD) were estimated using GIBBSF90+ software. The additive genetic variance ranged from 10.48 (0.60) to 24.12 (0.66), herd-year-season variance was 10.34 (0.40), herd-pen-milking date variance ranged from 4.91 (0.02) to 4.96 (0.02), permanent environmental variance ranged from 10.65 (0.44) to 16.94 (0.30), and residual variance ranged from 11.81 (0.01) to 14.60 (0.01). Heritability estimates ranged from 0.21 (0.01) to 0.47 (0.01), and repeatability estimates ranged from 0.51 (0.01) to 0.71 (0.01), and mean reliability of sires' breeding value predictions ranged from 0.81 to 0.89. Although caution is needed when disentangling associations between genetic effects, permanent environmental effects, and herd-pen-milking date contemporary groups, our results suggest that using daily milk weights and pen locations may improve the precision of genetic evaluations through increased sire PTA reliabilities for milk production traits in dairy cattle.