Identificación automática por resonancia magnética nuclear de la uva de origen en vinos blancos de Galicia

Resumo

A methodology for fast and automatic identification of the grape variety in white wines from Galicia based on Nuclear Magnetic Resonance spectroscopy (NMR) of intact wine samples is presented. The analysis of a collection of wines by metabolomic methods allowed the identification in the NMR spectrum of regions and peaks that are specific for each grape variety. In this way, the spectral “fingerprint” of monovarietals from Albariño, Godello, Treixadura and Palomino and the corresponding multivarietals where the minor component is at least 25%, were identified. The validity of these fingerprints applied to the fast identification of grape varieties used to vinification has been tested with 94 wines, showing a level of successful classification equal or better than 95%. The application of this methodology is straightforward and automatable, it only requires the measurement of the 1H NMR spectrum of a volume of 0,5 ml of intact wine and comparison by means of statistical metabolomic methodology (PCA-DA: Principal Component Analysis-Discriminant Analysis) of the spectrum with the fingerprint previously determined for each type of wine. The total time required for the wine classification is 10-15 min per sample