This application note describes a non-targeted metabolomic analysis approach to the classification of wine varieties, employing LC/QTOF MS and MassHunter Workstation software. Molecular feature extraction, data filtering, and statistical analysis utilizing ANOVA and PCA identified 26 marker compounds that were then used to build a PLSDA prediction model. An overall accuracy of 95.6% in discriminating between Cabernet Sauvignon, Merlot and Pinot Noir red wine varieties was obtained using the model. This approach may be widely applicable to the analysis and characterization of foods.
The document details an analytical study employing liquid chromatography coupled to quadrupole time-of-flight mass spectrometry (LC/QTOF MS) to classify red wine varieties. The methodology utilizes the Agilent Series 1200 SL Rapid Resolution LC system coupled with an Agilent 6530 accurate-mass Q-TOF MS.
The predictive model achieved an overall accuracy of 95.6% in discriminating between these three red wine varieties.
The analysis identified 26 marker compounds that were used to construct the classification model.
The column used for the analysis has a particle size of 1.8 pm.