Evaluation and quantification of associations between commonly suggested milk biomarkers and the proportion of grassland-based feeds in the diets of dairy cows

This study is a first step approach towards the prediction of the proportion of grassland-based feeds (%GB) in dairy cow diets with the aid of three different groups of milk biomarkers. We aimed to evaluate and quantify the associations between biomarkers commonly suggested in the literature and %GB in individual cows as a hypothesis-generating stage for the prospective establishment of accurate %GB prediction models. Consumers and governments financially encourage sustainable, local milk production making grass-based feeding, in grassland-dominated regions, of major interest. Milk from grassland-fed cows differs from that of other feeding systems by inferential fatty acids (FA), β-carotene content and yellow color; however, these biomarkers have not been evaluated together for their association with %GB. Using approved methods of parametric regression analysis, gas chromatography (GC), mid-infrared spectra (MIR) and color spectroscopy, we aimed to develop a first step towards an easy-to-implement, cost-effective milk-based control to estimate %GB in dairy cow diets. The underlying database was generated with 24 cows each fed one of 24 different diets gradually increasing in grass silage and decreasing in corn silage. Our results indicate that GC-measured α-linolenic acid, total n-3 FA and the n-6:n-3 ratio, MIR-estimated PUFA and milk red-green color index a* are robust milk biomarkers for constructing accurate prediction models to determine %GB. Based on simplified regression analysis, diets containing 75% GB should contain ≥ 0.669 and 0.852 g α-linolenic acid and total n-3 FA per 100 g total FA, respectively, and an n-6:n-3 FA ratio of