Hu Jian-ning, Fan Wen-lai, Xu Yan, Li Ji-ming, Yu Ying, Jiang Wen-guang
Thirty-two Cabernet Sauvignon wines from four different regions(Yantai,Xinjiang,Ningxia and Beijing) were analyzed using stir bar sorptive extraction(SBSE) in combination with gas chromatography-mass spectrometry(GC-MS) and liquid-liquid microextraction(LLME) coupled with gas chromatography-mass spectrometry(GC-MS),and 59 volatile compounds were quantified.Quantitative data were processed by multivariate data analysis,13 representative compounds(4-terpineol,1-nonanol,2-phenylethyl alcohol,2-phenylethyl acetate,3-methylthio-1-propanol,benzoic acid,4-methylphenol,α-terpineol,decanoic acid,ethyl octanoate,3-octanol,2-nonanol,ethyl hexadecanoate) were identified as potential predictors for the different geographical origins through analysis of variance(ANOVA) and principal component analysis(PCA).Then the geographical origin identification model of Cabernet Sauvignon wines was established using discriminant analysis(DA).In this model all the wines were successfully classified according to the geographical origins.Moreover,in order to evaluate the recognition ability of the model,cross validation was done and the results showed that the prediction accuracy for all the unknown wine samples reached 100%.