Prediction of the wine sensory properties related to grape variety from dynamic-headspace gas chromatography-olfactometry data

Wine extracts obtained by a dynamic headspace sampling technique were studied by quantitative gas chromatography-olfactometry (GC-O) to determine the aroma profiles of six young monovarietal Spanish white wines. A partial least-square regression study was carried out to look for models relating wine aroma properties with GC-O scores. Models were validated by sensory analysis, Four out of the five most important sensory descriptors were satisfactorily described by a model, and sensory tests confirmed most of the predictions. The main aroma differences between these wines are due to the ratio linalool/3-mercaptohexyl acetate. Floral, sweet, and muscat are positively related to the concentration of linalool and negatively to that of 3-mercaptohexyl acetate. Tropical fruit is related to the wine content in this last odorant. 2-Phenyl acetate, reinforced by other acetates, can also contribute to floral and sweet notes. Alkyl-methoxypyrazines lessen the tropical fruit note, and acetic acid lessens the muscat nuance.