Meat Authentication Based on Animal Species and Other Quality Meat Attributes (Protected Geographical Indication, Organic Production, and Halal and Kosher Products) by HPLC–UV Fingerprinting and Chemometrics

Abstract A simple and economic high-performance liquid chromatography with UV–vis detection (HPLC–UV) metabolomic fingerprinting methodology was developed and applied after a water extraction procedure to obtain sample chemical descriptors suitable for meat authentication by chemometrics. Three hundred meat samples involving different species (lamb, beef, pork, rabbit, quail, chicken, turkey, and duck) as well as different non-genetic attributes (protected geographical indications, organic production, and Halal and Kosher meats) were analyzed, and the obtained HPLC–UV fingerprints subjected to PCA and PLS-DA for classification and authentication. Excellent PLS-DA discrimination and classification performance was accomplished for calibration and cross-validation, with sensitivity and specificity values higher than 100% and 99.3%, respectively, and classification errors below 0.4%, when meat species were considered. The prediction capability when employing a classification decision tree consisting on consecutive dual PLS-DA models built using a hierarchical model builder was of 100% accuracy when 48 meat samples were subjected to the model as unknown samples. Multiclass PLS-DA classification performances when addressing meat geographical origin, organic productions and Halal and Kosher products were also very acceptable, with overall sensitivity and specificity values higher than 91.2%, and classification errors below 6.9%. Finally, fraudulent meat adulteration cases involving PGI, organic and Halal and Kosher adulterated meats were evaluated by partial least squares (PLS) regression, allowing the detection and quantitation of adulteration levels within the range from 15 to 85% with prediction errors below 6.6%, demonstrating the suitability of the proposed methodology to assess meat authenticity.