Wines: The chemical signature of fine wines has been identified through artificial intelligence

According to a study. Researchers have managed to trace the "exact origin" of several Bordeaux Grand Crus solely based on their chemical components, detected through AI.

This method, presented in the journal Communications Chemistry, could help "combat counterfeiting" and "guide decision-making for wine producers" in the context of climate change, as stated in a press release by the University of Geneva, which conducted the study in partnership with the University of Bordeaux.

In total, researchers analyzed 80 red wines from seven estates in the Bordeaux vineyard, separating their molecules—extremely numerous due to the complexity of blends—using the classical method of gas chromatography, regularly used for wine analysis.

A detector—the mass spectrometer—then translates the chemical separations and molecule density into a graph, resembling a gigantic electrocardiogram. The result can include up to 30,000 points, making detailed analysis tedious.

The team from the Department of Neuroscience at the University of Geneva, led by Alexandre Pouget, developed a processing method for this raw data using "machine learning," an AI automatic learning system.

The method allowed the researchers to discern, among these thousands of points, a similar arrangement on vintages from the same estate, demonstrating that each terroir has a "unique chemical signature," according to the researchers.

Still using AI, by reducing the data obtained for each castle into a simple graph with axes and ordinates, one can find "a physical map of Bordeaux" with "the same location of the castles" and "the same distance between them" as on a geographical map.

According to Mr. Pouget, this identification technique, costing around a hundred euros, could easily detect counterfeits or, in the future, help winemakers find the right blend of grape varieties and plots to replicate the exact signature of a wine from their estate.

According to Stéphanie Marchand, a professor at the Institute of Vine and Wine Sciences at the University of Bordeaux and co-author of the study, this method allows for "a better understanding of commonalities between wines from the same terroir."

In "a period where terroirs are mistreated by climate change and evolving practices," the obtained data will help preserve the identity and expression of a wine and its terroir, she notes.