Assessing Serra da Estrela PDO cheeses’ origin-production date using fatty acids profiles

• Published in: Journal of food measurement & characterization 2019-12, Vol.13 (4), p.2988-2997
• Main Authors: Reis Lima, M. J., Bahri, Hamdi, Sá Morais, Jorge, Veloso, Ana C. A., Fontes, Luísa, Lemos, Edite T., Peres, António M.
• Format: Article
• Language: English
• Subjects: Algorithms; Biomarkers; Cheese; Chemical composition; Chemistry; Chemistry and Materials Science; Chemistry/Food Science; Computer simulation; Consumers; Dairy products; Engineering; Fatty acids; Fish oils; Food Science; Manufacturing; Organic chemistry; Original Paper; Simulated annealing
• ISSN: 2193-4126; 2193-4134
• DOI: 10.1007/s11694-019-00219-z

Serra da Estrela is a Portuguese traditional cheese produced with raw ewe’s milk from “Churra Mondegueira” and “Bordaleira” autochthonous breeds and the wild thistle flower ( Cynara cardunculus L.), which benefits from the status of Protected Designation of Origin. Cheese chemical composition, namely the fatty acids profile, depends on milk composition and on manufacturing practices. Thus, the identification of possible chemical biomarkers capable of classifying Serra da Estrela cheeses according to the dairy manufacturing plant, geographical origin or production date would be of utmost relevance for producers and consumers. A typical fatty acids profile, including 23 saturated and unsaturated fatty acids, was identified for the studied cheeses, being butyric, caproic, caprilic, capric, lauric, miristic, palmitic, stearic, oleic, linoleic and its trans-isomer and α-linolenic acids the most abundant ones (relative mean abundances ranging from 1.4% ± 0.5% to 23.9% ± 1.9%). Linear discriminant models were established based on the most discriminative fatty acids (namely, caproic, caprilic, undecanoic, lauric, pentadecanoic, palmitic, palmitoleic, heptadecanoic, oleic, linoleic trans-isomer, heneicosanoic and arachidonic acids) that included less abundant fatty acids, which were selected using the simulated annealing algorithm. The established models enabled assessing cheeses’ origin (models based on 10–12 fatty acids) and/or production date (model based on 20 fatty acids) with predictive sensitivities of 71–88%. Therefore, fatty acids profiles coupled with chemometric techniques, could be foreseen as a fingerprint of cheese’s genuineness, enhancing the consumers’ confidence when purchasing this high-value cheese.