Chemometric analysis of milk farming systems by fatty acid profile

This study investigates the use of multivariate analysis of milk fatty acid (FA) patterns to authenticate milk from different intensive dairy farming systems, characterized by varying forage-to-concentrate (F/C) ratios (70:30 or 60:40) and forage types (hay or maize silage). A total of 72 milk samples were collected: 50 from dairy farms, grouped according to the type and amount of forage in the cows’ diets, and 22 retail samples labeled as “High Quality.” A chemometric model was developed using linear discriminant analysis (LDA) with leave-one-out cross-validation. Key discriminant fatty acids such as C8:0 (caprylic acid), C12:1(lauroleic acid), iso C14:0 (isomyristic acid), C15:0 ante iso (12-methyltetradecanoic acid), C16:0 iso (14-methylpentadecanoic acid), anteiso C17:0 (14-methylhexadecanoic), C18:1 ω-7 (cis-vaccenico), C18:3 ω-6 (γ− linolenic acid), C20:5 ω-3 (eicosapentaenoic acid), C20:2 ω-3 (eicosadienoic acid), and C20:0 (arachidic acid), enabled a correct classification rate of 97 % based on the farming system. To validate the LDA model and assess its applicability to milk from grazing systems (F/C = 100), a blind test was conducted using 46 independent milk samples. The model achieved a classification accuracy of 95 %. Our results suggest that milk fatty acid patterns can be a reliable chemical biomarker for determining the milk’s origin.