LC-MS based metabolomics for the authentication of selected Greek white wines

LC-MS based metabolomics provide a new perspective in wine authentication enabling a thorough investigation of its chemical composition. In this study, 97 monovarietal white wines derived from four indigenous Greek grape varieties (Assyrtiko, Moschofilero, Malazouzia and Savatiano) and produced in two PDO and one PGI winemaking regions were analyzed with the use of ultra-high-performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UPLC-QTOF/MS). A targeted metabolomics method was developed, validated and applied to the analyzed wine samples for the identification and quantification of 22 metabolites. Another 79 compounds were tentatively identified using a smart suspect screening workflow based on an in-house developed database. Then, the random forest algorithm was applied on the normalized peak areas of both target and suspect data for supervised statistical analysis. A robust classification model was built enabling the classification and prediction of the varietal origin of the wine samples in acceptable levels. Specific biomarkers, contributing significantly to the discrimination of wine variety, were recognized.

Automated summary

LC-MS-based metabolomics was used to analyze white wines from four Greek grape varieties, developing a targeted method for the identification and quantification of 22 metabolites and tentatively identifying 79 compounds. A robust classification model was built using the random forest algorithm for the discrimination of wine variety.