Advanced multimodeling of isotopic and elemental content for cheese authentication

BACKGROUND Dairy product authentication is a key topic in food science because of concerns about food fraud, traceability and consumer protection. Cheese, as a complex matrix, reflects environmental and technological influences, enabling origin and type differentiation. Different chemometric models, such as linear discriminant analysis (LDA) and generalized linear models (GLMs) were developed in order (i) to discriminate cheese samples according to their geographical origin and cheese type and (ii) to evaluate of interaction between two main effects (geographic and type of cheese), based on elemental and isotopic compositions (Mg, K, Ca, Cr, Mn, Co, Ni, Cu, Zn, As, Cd, Pb, δ 13 C cheese and δ 13 C casein ). RESULTS The main predictors for differentiation of cheeses geographical origin obtained within LDA model were Mn, Zn, Co, Ca, Cr and δ 13 C cheese . An overall initial classification of 82.2% was obtained, whereas, in the cross‐validation procedure, a lower percentage was achieved. Cu, Zn, Mg, and δ 13 C casein were identified as significant parameters for the differentiation of cheese varieties. GLMs were developed for each variable (Mg, K, Ca, Cr, Mn, Co, Ni, Cu, Zn, As, Cd, Pb, δ 13 C casein and δ 13 C cheese ) to assess the main effects for geographical origin and cheeses type, as well as their interaction. CONCLUSION In an investigation of different types of cheese through elemental content and carbon isotopic ratio ( 13 C/ 12 C) of bulk cheeses and extracted casein, LDA and GLMs were suitable for the development of differentiation models of cheese samples according to their geographical origin and cheese type, as well as to evaluate the interaction between two main effects (geographic and type of cheese).