Derivation and genome-wide association study of a principal component-based measure of heat tolerance in dairy cattle

Heat stress represents a key factor that negatively affects the productive and reproductive performance of farm animals. In the present work, a new measure of
tolerance to heat stress for dairy cattle was developed using principal component analysis. Data were from 590,174 test-day records for milk yield, fat and protein
percentages, and somatic cell score of 39,261 Italian Holstein cows. Test-day records adjusted for main systematic factors were grouped into 11 temperature humidity
index (THI) classes. Daughter trait deviations (DTD) were calculated for 1,540 bulls as means of the adjusted test-day records for each THI class. Principal
component analysis was performed on the DTD for each bull. The first 2 principal components (PC) explained 42 to 51% of the total variance of the system across the
4 traits. The first PC, a measure of the level at which the curve is located, was interpreted as a measure of the level at which the DTD curve was located. The second
PC, which shows the slope of increasing or decreases DTD curves, synthesized the behavior of the DTD pattern. Heritability of the 2 component scores was moderate
to high for level across all traits (range = 0.23–0.82) and low to moderate for slope (range = 0.16–0.28). For each trait, phenotypic and genetic correlations between
level and slope were equal to zero. A genome-wide association analysis was carried out on a subsample of 423 bulls genotyped with the Illumina 50K bovine bead
chip (Illumina, San Diego, CA). Two single nucleotide polymorphisms were significantly associated with slope for milk yield, 4 with level for fat percentage, and 2 with level and slope of protein percentage, respectively. The gene discovery was carried out considering windows of 0.5 Mb surrounding the significant markers and
highlighted some interesting candidate genes. Some of them have been already associated with the mechanism of heat tolerance as the heat shock transcription factor
(HSF1) and the malonyl-CoA-acyl carrier protein transacylase (MCAT). The 2 PC were able to describe the overall level and the slope of response of milk production
traits across increasing levels of THI index. Moreover, they exhibited genetic variability and were genetically uncorrelated. These features suggest their
use as measures of thermotolerance in dairy cattle breeding schemes.