Abstract #367

# 367
Crossbreeding affects the production performance of dairy cows exposed to a range of temperature and humidity in a pasture-based system.
J. Graham*1, S. Biffani2, F. Tiezzi1, 1Department of Animal Science, North Carolina State University, Raleigh, NC, 2Istituto di Biologia e Biotecnologia Agraria, Milan, Italy.

The aim of this study was to evaluate the response in production traits of different crossbred combinations exposed to increasing temperature and humidity. Data included 15,790 test-day records from purebred and crossbred cows (n = 713) located at the North Carolina Department of Agriculture Research Station (Goldsboro, NC) from 1996 to 2018. Traits analyzed were: milk yield (MY), fat percentage (F%), protein percentage (P%), somatic cell score (SCS), fat kilogram (Fkg), protein kilogram (Pkg), total somatic cell score (TotSCS) and fat to protein ratio (F:P). The 8 traits were analyzed by fitting a mixed model and a random regression model, with heterogeneous residual variances as defined by the environmental descriptor. Fixed effects in the mixed model included stage of lactation, parity, environmental descriptor (E), breed group (BG) and the interaction between E and BG (E×BG). Random effects included test date and animal. BG consisted of 8 different Holstein (H) and Jersey (J) combinations: H×H, J×J, H×J, J×H, J×JH, H×JH, J×HJ, H×HJ. Environmental descriptors were evenly distributed into 5 classes (class 5 being the most challenging). E×BG estimates were used to determine reranking of breed groups when exposed to more challenging environments. BG that reranked between extreme classes of E were: H×H vs H×J for MY, H×JH vs H×J for Pkg. Fixed effects in the random regression model consisted of stage of lactation, parity and zero, first and second-order Legendre polynomials on E nested into the random effects of breed (with breed relationship matrix), additive genetic (with pedigree relationship matrix) and permanent environmental. The random effect of breed had larger variance explained than additive genetic for Pkg (intercept) and F%, Fkg, Pkg and F:P (slope). For the same traits, intercept and slope of the breed random effect showed similar amounts of variance explained, indicating a sizable breed by environment interaction. Mean heritability estimates over values of E were ~0.2 for MY, F%, P%, and ~0.1 for all others. Heritability decreased over values of E for the traits F%, P%, Fkg, Pkg, F:P, and increased for all others.