Untargeted lipidomics by LC-QTOF-MS for biomarker selection and discrimination of plant- and animal-derived rennet in cooked-curd ovine cheese at different ripening stages

In this study, the lipidomic alterations in cooked-curd ovine cheese were comprehensively assessed by comparing samples produced using either vegetal (Cynara cardunculus L.) or animal (calf) rennet. Lipid extracts were analyzed using high-resolution liquid chromatography-quadrupole time-of-flight mass spectrometry (LC-QTOF-MS), and the resulting data were processed through multivariate statistical analyses to explore compositional changes across 18 months of ripening. The use of cardoon rennet, led to distinct modifications in the lipid profile compared to conventional calf rennet. A broad range of lipid classes, including phosphatidylserines, triacylglycerols, free fatty acids, phosphatidylcholines, ceramides, sphingomyelins, phosphatidylethanolamines, and monoacylglycerols, were significantly affected by the type of rennet employed. In contrast, the ripening process mainly influenced the abundance and composition of free fatty acids, reflecting ongoing lipolysis and lipid remodeling. To enhance classification and biomarker discovery, a forward stepwise interval partial least squares discriminant analysis (PLS-DA) was applied, enabling the identification of two optimal feature subsets for discriminating samples based on rennet type and ripening stage. The resulting models achieved high classification accuracies of 88 % for rennet type (cardoon vs. calf) and 91 % for ripening time (24 h, 12 months, 18 months), underscoring the strong lipidomic signature linked to both technological variables. These results allowed to better understand how the type of rennet and the ripening process can change the composition of lipids in cheese. This knowledge can be useful for improving product quality, ensuring authenticity, and developing new types of traditional cheeses with specific features.