Genetic variation of predicted milk fatty acids groups in Canadian Holsteins

Abstract Text:

The objective of this study was to investigate genetic variability of mid infra-red predicted fatty acids groups in Canadian Holstein cows. Milk samples were collected by CanWest DHI (Guelph, ON, Canada) and Valacta (Sainte-Anne-de-Bellevue, QC, Canada) during routine milk recordings. Milk samples were analyzed using MilkoScan FT6000 spectrometers (Foss, Hillerød, Denmark). Milk mid infra-red spectra generated from January 2013 to July 2015 were standardized and then predicted for five groups of fatty acids: short-chain (C₄-C₁₀), medium-chain (C₁₂-C₁₆), long-chain (C₁₇-C₂₂), saturated (no double bond), and unsaturated (one or more double bonds) fatty acids. The predicted fatty acid values were log transformed in order to improve normality. The data set included 49,127 test-day records from 10,029 first lactation Holstein cows in 810 herds. The total number of animals in the pedigree was 76,074. The random regression animal test-day model included: days in milk, herd test day, and season-age of calving (polynomial regression) as fixed effects, and herd-year of calving, animal and permanent environment effects as random polynomial regressions, and  random residual effect. The significance of fixed effects and the best degree of the fixed Legendre polynomial regressions for season-age effect (3^rd degree) were determined using AI-REML. Bayesian methods with Gibbs sampling were then used for fitting models with different degree of random regressions, assuming  the best degree for fixed regressions, and the same increasing degree for all random effects (from intercept only to 4^thdegree). Fourth degree random regressions yielded the best fitting based on the Deviance Information Criterion (DIC). No polynomials with degree higher than 4 were fit due to low number of cows with more than 5 fatty acid measurements and the cubic shape of the phenotypic distribution of the fatty acid groups. The estimate of average daily heritability over the lactation for medium-chain fatty acid (0.37) was higher than for short-chain (0.30) and long-chain (0.24). The average daily heritability for saturated fatty acid (0.38) was larger than for unsaturated fatty acid group (0.23). These results provide evidence for the existence of genetic variation in fatty acids groups, and thus indicate possibility of altering milk fatty acid composition through genetic selection.

Keywords: milk fatty acids, mid infra-red, random regression model