A multivariate approach to exploring interrelationships among milk fatty acids across ruminant species

The milk fatty acid (FA) profile is influenced by complex interactions among animal species, physiology, feeding, and management, resulting in modification to the nutritional and functional properties of milk fat. This study aimed to identify latent biological and nutritional factors influencing milk FA composition in ruminants using multivariate factor analysis (MFA), applied across 4 species: dairy cows (Holstein-Friesian and Brown Swiss), sheep (Sarda and Massese), Mediterranean buffaloes, and goats (Saanen). A total of 1,960 mid-lactation animals (150 ± 16 DIM for cows; 90 ± 5 DIM for sheep; 162 ± 14 DIM for buffalo; 86 ± 3 DIM for goats), raised under commercial farm conditions and subjected to diverse feeding regimens, were included in the study. All milk samples were analyzed by the same laboratory using GC under consistent analytical conditions. Factor analysis was carried out on 49 individual FA. An ANOVA on the factor scores was performed to assess the effects of species and feeding strategies, specifically lipid supplementation in cow milk and n-3 supplementation in sheep milk. The MFA was able to extract 7 latent factors with specific biologic meaning: synthesis of odd- and branched-chain fatty acids from ruminal bacteria (factor 1: “OBCFA”), de novo FA synthesis via acetate in mammary gland (factor 2: “Mammary activity”), biohydrogenation of linoleic acid (factor 3: “BH_1”), milk fluidity regulation (factor 4: “Milk fluidity”), mitochondrial Beta-oxidation of FA (factor 5: “Beta-oxidation”), biohydrogenation via vaccenic acid (factor 6: “BH_2”), ruminal 16:1c7 elongation (factor 7: “16:1c7 elongation”), regulation of branched FA, n-6 metabolism (factor 10: “n-6 FA”), and synthesis of short-chain FA (factor 11: “SCFA”). According to a previous study on MFA, a variable was considered to be associated with a specific factor if the absolute value of its correlation with the factor was ≥0.60. Several factors reflected established metabolic processes, such as “OBCFA,” “Mammary activity,” and “BH_1.” Others, such as “Beta-oxidation” and “16:1c7 elongation,” captured novel pathways not previously described in singlespecies analyses. Importantly, this multispecies MFA approach revealed both conserved and species-specific metabolic signatures. For example, whereas core factors such as “SCFA” and “OCFA” appeared across all species and aligned with prior bovine-focused studies, certain factors (e.g., “n-6 FA” and “BCFA”) showed species-dependent variation, likely reflecting differences in animal diet and ruminal microbial ecosystems. The extracted factor scores were useful to evaluate the effects of species (cow, goat, sheep, and buffalo) and different feeding regimens (applying the scores only to Holstein- Friesian cows for lipid supplementation and Sarda sheep for n-3 source supplementation). This approach allowed us to assess the influence of these dietary interventions on the latent factors affecting milk fatty acid composition. The resulting factors offer biologically meaningful, synthetic variables that can be used for future phenotypic, genetic, or nutritional modeling. Moreover, this approach enhances our understanding of mammary lipid metabolism across species and underscores the value of integrative multivariate methods in dairy science. These findings have implications for animal selection, feeding practices, and management strategies aimed at improving the nutritional quality of milk fat.