The milk flavor can be attributed to the presence of numerous flavor molecules and precursors. In this study, we employed widely targeted metabolomic analysis techniques to analyze the metabolic profiles of various milk samples obtained from goats, sheep, dairy cows, and buffaloes. A total of 631 metabolites were identified in the milk samples, which were further categorized into 16 distinct classes. Principal component analysis (PCA) suggested that the metabolite profiles of samples from the same species exhibit clustering, while separated patterns of metabolite profiles are observed across goat, sheep, cow, and buffalo species. The differential metabolites between the groups of each species were screened based on fold change and variable importance in projection (VIP) values. Five core differential metabolites were subsequently identified, including 3-(3-hydroxyphenyl)-3-hydroxypropanoic acid, inosine 5'-triphosphate, methylcysteine, N-cinnamylglycine, and small peptide (L-tyrosine-L-aspartate). Through multiple comparisons, we also screened biomarkers of each type of milk. Our metabolomic data showed significant inter-species differences in the composition and concentration of some compounds, such as organic acids, amino acids, sugars, nucleotides, and their derivatives, which may affect the overall flavor properties of the milk sample. These findings provided insights into the molecular basis underlying inter-species variations in milk flavor.
Keywords: UPLC–ESI–MS/MS-based metabolomics; flavor profile; goat; milk.