Multivariate statistical assessment of regional differences in mineral composition of cow’s milk and feed in Kosovo

This study investigated regional differences in the composition of raw cow’s milk and feed to determine their interrelationship using multivariate analysis. The study included 104 farms producing raw cow’s milk (with more than 10 dairy cows) across seven regions of Kosovo during the 2022-2024 period. Physicochemical parameters of milk (pH, density, acidity, fat, lactose, protein, water activity (aw), solids-not-fat (SNF), conductivity, FP and ash) were analyzed, while macroelements (Ca, K, P, Mg, Na) and microelements (Fe, Zn, Cu, Mn) were determined in both milk and feed. Statistical analysis included descriptive statistics and one-way ANOVA (with Tukey’s HSD post-hoc test) to determine regional differences. Principal component analysis (PCA) was applied to investigate multivariate effects and relationships between milk and feed parameters. Statistically significant regional differences were confirmed in most of the observed physicochemical and mineral parameters of raw cow’s milk. The average mineral composition of milk from Kosovo dairy farms showed that higher concentrations of phosphorus (P) and sodium (Na) were observed relative to the internationally recommended values (FAO), while the concentrations of Ca, K, Mg, Zn and Mn were lower than recommended. A positive correlation was observed between Mg and Ca in animal feed and milk density, which was confirmed by Pearson correlation coefficients rCa= 0.787 and rMg=0.767. The application of multivariate analysis (MVA) was extremely important to obtain an integrated assessment of nutrient transfer and regional differentiation. Overall, the findings highlight that improving feed formulation and mineral balance could improve the nutritional quality of milk, provide a scientific basis for regional dairy product branding initiatives and promote sustainable livestock management.