CLASSIFICATION OF MENDOZA (ARGENTINA) RED WINES BY ICP-MS MULTI-ELEMENT ANALYSIS AND MULTIVARIATE STATISTIC TECHNIQUES

CANIZO, BRENDA; BRUSA, LUCILA; SIGRIST, MIRNA; PELLERANO, ROBERTO G.; WUILLOUD, RODOLFO G.

Mendoza

Simposio; 15th RIO SYMPOSIUM ON ATOMIC SPECTROMETRY; 2019

Universidad Nacional de Cuyo

Wine geographical origin authentication has been an important concern in global wine trade. Specially, the authenticity of the production area in which the interactions between the identifiable physical and biological environment and the vitivinicultural practices used provide distinctive characteristics of wine.1 Regional classification using mineral profile with chemometrics has been worldwide explored for wines. The aim of this work was to investigate the potential use of several multivariate statistics tools combined with multielemental data obtained by ICP-MS analysis for the geographical differentiation of wines from Mendoza from others wine region (Salta and La Rioja). One hundred wines were obtained with the collaboration of the National Institute of Agricultural Technology (INTA). The samples were acid digested with ultrapure nitric acid and analyzed. All samples were characterized by 27 descriptors: Ag, Al, As, Ba, Be, Cd, Co, Cr, Cs, Cu, Fe, Ga, Hg, Li, Mn, Mo, Ni, Pb, Pd, Rb, Sb, Se, Sn, Sr, Tl, V and Zn. Basic chemometrical characterization was made by PCA. The results showed that the samples from Mendoza region could be differentiated from wines of Salta and La Rioja regions. Moreover, three chemometric models were selected and tested, Logistic Regression (LR), Random Forest (RF) and k-nearest neighbour (K-NN) and displayed similar degrees of success in the prediction of test samples. The order of successful identification rates was as follows: LR>RF>k-NN. The use of a linear method resolved the classification problem, being the Logistic Regression the ideal model for discriminating wine samples of Mendoza from the others regions, with an overall classification value of AUC of 97.2%. LR method is proved to be a promising and simple tool in wine classification analysis and quality control to the wine-making industry. Acknowledgements CONICET, UNCUYO and FONCyT-ANPCyT.References [1] Pereira, L., Gomes, S., Barrias, S., Gomes, E., Baleiras-Couto, M., Fernandes, J., Martins-Lopes, P., Beverages 4 (2018) 1-21.