Advancing specific local wine identification through integration of ICP-MS mineral wine profiles and Machine Learning: A case study on discrimination among the ten Beaujolais crus
Abstract
The Beaujolais vineyard, located in France's Auvergne-Rhône-Alpes region, is internationally recognized for its geological diversity and the strong identity of its ten crus. Guaranteeing authenticity of these appellations is essential for producers and regulators, as mislabeling and fraud can undermine both economic value and consumer trust. In this study, we introduce the Mineral Wine Profile (MWP), a high-throughput ICP-MS protocol enabling the semi-quantitative analysis of more than 40 mineral elements. A total of 1288 Gamay wines from the 2022 and 2023 vintages, collected across the ten Beaujolais crus, were analyzed. Multivariate statistical methods and machine learning were applied to assess their discriminatory power. Principal Component Analysis clearly separated vintages, while Random Forest models achieved 98% accuracy for vintage classification. Mixture Discriminant Analysis further differentiated the ten crus with an average accuracy of 92%. Misclassifications mostly occurred between geographically neighboring crus, reflecting overlapping soil structures and terroir continuity. Element-soil correlations (e.g., Fe enriched in clay-rich soils, Mn in sandy soils, Cs in silty soils) provided mechanistic insights into terroir differentiation. These results confirm that wine mineral fingerprints can act as reliable chemical memories of origin. By combining elemental profiling with Machine Learning, this work demonstrates a robust, reproducible, and scalable approach for intra-regional wine authentication. Beyond its scientific contribution, the MWP approach offers practical applications for the valorization of terroirs, protection of Appellations d'Origine Contrôlée (AOC), and fraud prevention in the wine industry.
Related
- Advancing Marine Mammal Conservation Through Omics Tools and Integration of Indigenous Ecological Knowledge
- Advancing Human Skin Equivalents: The Crucial Role of Neurovascular Integration
- Perceptual integration of multisensory haptic, visual, and auditory feedback for roughness discrimination in augmented reality
- Advancing Spatial Pattern Integration in Species Distribution Modelling With Computer Vision: How Grain Size and Extent Concepts Shape Predictions of Emerging Aquatic Insects
- Protein novelty, sequence similarity, and negative sampling controls reveal benchmark specific discrimination in protein protein interactionprediction
BibTeX
@article{clement2026advancing,
title = {{Advancing specific local wine identification through integration of ICP-MS mineral wine profiles and Machine Learning: A case study on discrimination among the ten Beaujolais crus}},
author = {Yohann Clément and Pierre Lantéri and Agnès Hagège and Bertrand Chatelet and Théodore Tillement and Coraline Duroux and Leticia Sarlo and François Lux and Olivier Tillement},
journal = {Talanta},
year = {2026},
month = oct,
doi = {10.1016/j.talanta.2026.129581},
url = {https://hal.science/hal-05796350v1},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Advancing specific local wine identification through integration of ICP-MS mineral wine profiles and Machine Learning: A case study on discrimination among the ten Beaujolais crus
AU - Yohann Clément
AU - Pierre Lantéri
AU - Agnès Hagège
AU - Bertrand Chatelet
AU - Théodore Tillement
AU - Coraline Duroux
AU - Leticia Sarlo
AU - François Lux
AU - Olivier Tillement
PY - 2026
DA - 2026/10/29
JO - Talanta
DO - 10.1016/j.talanta.2026.129581
UR - https://hal.science/hal-05796350v1
AB - The Beaujolais vineyard, located in France's Auvergne-Rhône-Alpes region, is internationally recognized for its geological diversity and the strong identity of its ten crus. Guaranteeing authenticity of these appellations is essential for producers and regulators, as mislabeling and fraud can undermine both economic value and consumer trust. In this study, we introduce the Mineral Wine Profile (MWP), a high-throughput ICP-MS protocol enabling the semi-quantitative analysis of more than 40 mineral elements. A total of 1288 Gamay wines from the 2022 and 2023 vintages, collected across the ten Beaujolais crus, were analyzed. Multivariate statistical methods and machine learning were applied to assess their discriminatory power. Principal Component Analysis clearly separated vintages, while Random Forest models achieved 98% accuracy for vintage classification. Mixture Discriminant Analysis further differentiated the ten crus with an average accuracy of 92%. Misclassifications mostly occurred between geographically neighboring crus, reflecting overlapping soil structures and terroir continuity. Element-soil correlations (e.g., Fe enriched in clay-rich soils, Mn in sandy soils, Cs in silty soils) provided mechanistic insights into terroir differentiation. These results confirm that wine mineral fingerprints can act as reliable chemical memories of origin. By combining elemental profiling with Machine Learning, this work demonstrates a robust, reproducible, and scalable approach for intra-regional wine authentication. Beyond its scientific contribution, the MWP approach offers practical applications for the valorization of terroirs, protection of Appellations d'Origine Contrôlée (AOC), and fraud prevention in the wine industry.
SN - 0039-9140
ER -