Origin verification of French red wines using isotope and elemental analyses coupled with chemometrics. (1st March 2021)

• Record Type: Journal Article
• Authors: Wu, Hao; Lin, Guanghui; Tian, Ling; Yan, Zhi; Yi, Bingqing; Bian, Xuehai; Jin, Baohui; Xie, Liqi; Zhou, Haichao; Rogers, Karyne M.
• Is Part Of: Food chemistry. Volume 339(2021)
• Publication Date: 2021-03-01
• Subjects: Trace elements -- Stable isotopes -- French wine -- Geographical origin -- Fraud detection; Food -- Analysis -- Periodicals; Food -- Composition -- Periodicals; 664
• DOI: 10.1016/j.foodchem.2020.127760
• Languages: English
• ISSNs: 0308-8146

Abstract

Highlights: French red wine was classified to verify origin. We used stable isotopes and elemental analyses combined with chemometric methods. The most important variables for traceability were Mg, Mn, Na, Sr, Ti and Rb. Geographical separation of wine was high with an accuracy up to 98.2% using ANN. Traceability was possible on a regional and sub-regional scale. Traceability was not possible using isotopic and elemental analysis at a village scale. Abstract: Origin verification of 240 French wines from four regions of France was undertaken using isotope and elemental analyses. Our aim was to identify and differentiate the geographical origin of these red wines, and more importantly, to build a classification tool that can be used to verify geographic origin of French red wines using machine learning models. Multivariate analyses of the isotopic and elemental data revealed that it is possible to determine the geographical origin of French wines with a high level of confidence for most regions analyzed in this study. The wine verification accuracy of four French wine producing regions of Bordeaux, Burgundy, Languedoc-Roussillon and Rhone using an Artificial Neural Network (ANN) method was 98.2%. The results also show that ANN is more suitable than Discriminant Analysis for this verification purpose. The most important variables for French wine regional traceability were Mg, Mn, Na, Sr, Ti and Rb.