Development of genomic predictions for heat tolerance in US Holstein cattle

Heat stress (HS) represents a major threat to the overall welfare and performance of dairy cows; still, genetic evaluations for heat tolerance are not available to dairy producers in the United States. Zoetis has accumulated a large amount of data suitable for genetic and genomic evaluation for heat tolerance. Therefore, the objective of this study was to develop genomic predictions for heat tolerance in US Holstein cattle using milk yield and conception at first service based on producer-recorded data coupled with publicly available weather information. Test-day (TD) milk yields and insemination records were available from 370 herds distributed across 30 states, spanning a 20-yr period (2001-2021). Weather data were extracted from the National Oceanic and Atmospheric Administration database and used to calculate the daily temperature-humidity index (THI). Productive and reproductive records from each herd were merged with the THI data from the nearest weather station. After matching, about 83 million (M) TD milk records and over 6M inseminations were available for analysis. The effect of HS was modeled using a bivariate reaction norm linear model that assumed the negative impact of HS occurred at THI ≥70. The evaluation was conducted using the single-step genomic BLUP (ssGBLUP) methodology, applying the algorithm for proven and young animals. Over 2M genotyped animals were available. The model for conception at first service (CFS) included herd-year-season (HYS) of calving, parity, breeding type, and voluntary waiting period as fixed effects, whereas the additive genetic effect, the random regression on THI for HS genetic effect, the permanent environment, and the random regression on THI for the permanent environment effect were considered random. The model for milk yield included HYS, parity, and DIM classes as fixed effects and the same random effects as described for CFS. Heritabilities of the HS genetic components varied depending on THI, ranging from 0.17 to 0.47 for milk yield and from 0.08 to 0.55 for CFS as THI increased. Genetic correlations between additive genetic and HS-influenced genetic components were -0.29 and -0.05 for milk yield and CFS, respectively. To enable running evaluation in a commercial setting with very large number of genotypes within a reasonable time, the original threshold-linear model was changed to a bivariate linear model after performing a linear transformation of CFS. Two EBVs for heat tolerance were obtained, Milk_THI (expressed as change of daily milk yield in kg) and CFS_THI (expressed as the change of probability of conception) per unit increase of THI. Milk_THI ranged from -1.3 to 1.0 kg per day and unit of THI, and CFS_THI ranged from -6.2 to 5.3 percentage points. The range of EBV for both traits suggest potential to enhance heat tolerance in Holstein cows without negatively affecting milk production or fertility.