Abstract

The identification of grapevine (Vitis vinifera L.) genotypes is conventionally a laborious activity that must be carried out by specialized staff. In this work a novel and simple method for differentiation of grapevine genotypes is presented. Direct measurements of leaves by attenuated total reflectance Fourier-transform infrared spectroscopy (ATR-FTIR) combined with chemometric methods were used for classification of six genotypes (five varieties and a pair of clones), viz. Cinsault, Gewurztraminer (clone 643), Moscatel de Alejandría, País, Pinot Noir (French clone 777), Pinot Noir (local clone ‘Valdivieso’). These were successfully classified and identified through supervised pattern recognition methods such as soft independent modeling of class analogy (SIMCA) and partial least square discriminant analysis (PLS-DA). The error rate for spectra classification of test sets by both models was 0.08. The results demonstrate the advantages of using ATR-FTIR as a rapid and non-destructive tool that achieves accurate grapevine genotype differentiation.