Genetics of the mineral contents in bovine milk predicted by mid-infrared spectrometry

Knowing the contents of minerals in milk like Ca or Na could be interesting to improve the nutritional quality of milk and to assess the animal health status. This study had two aims: 1) development of mid-infrared equations for mineral contents in milk by using an approach combining multiple countries, breeds, and production systems and 2) study of the genetic variability of these traits in the Walloon Holstein dairy cattle. Samples included in the calibration set were collected in Belgium, Luxembourg and France over 5 years. The calibration set included at least 400 samples analyzed by coupled plasma atomic emission spectrometry to quantify the contents of Na, Ca, Mg, P and K. The calibration coefficient of determination ranged between 0.69 for K and 0.93 for Na. The standard error of cross-validation was 63.35, 49.24, 64.33, 7.04, and 93.22 mg/kg of milk for Na, Ca, P, Mg and K. From these results, the quantification of milk minerals by mid-infrared is feasible. These equations were applied to more than 140,000 spectral records collected from 43,797 first parity Holstein cows in 1,233 herds. The variance components were estimated using Gibbs Sampling using single trait random regression models derived from the one used for the Walloon genetic evaluation of milk production traits. First results gave a daily heritability of 0.26 for Na, 0.45 for Ca, 0.48 for P, 0.46 for Mg, and 0.41 for K. Moderate negative genetic correlations were found between Na and the other studied traits. The highest correlation (0.69) was observed between P and Mg. These results confirmed the genetic variability of these traits. Further studies will be conducted to study the relationship between these traits and other traits (e.g., production, health).