论文部分内容阅读
本文创新性地将SPME-GC-MS检测手段和代谢组学数据处理技术结合应用于白酒特征化合物鉴定和真假区分中。17个白酒样品经顶空固相微萃取,富集酒中挥发性化合物。所得GC-MS数据集经代谢组学技术降维处理,经聚类分析真酒和假酒样品得到了正确的区分,同时其他酒系列所得分类结果基本符合实际酒样信息。在主成分分析中,提取了PC1和PC2两个主成分,解释了不同酒样品特征变量58.7%的方差信息,其中PC1为44.7%,真假酒霍特林椭圆区域区分明显,可视化效果直观。其他系列酒样也保持了对真实酒样信息的吻合。利用偏最小二乘判别分析法建立酒类相关模型,去除基酒样品以突出真假酒的物质区别,得出变量重要性表并且查库得出相关物质十二种,差异最显著的前三位特征化合物是乙酸丁酯、己酸异戊酯和己酸-1-甲乙酯。
In this paper, the SPME-GC-MS detection method and the metabolomics data processing technology are applied to the identification of liquor characteristic compounds and the distinction between true and false. 17 liquor samples were subjected to headspace solid-phase microextraction to enrich the volatile compounds in the wine. The obtained GC-MS datasets were reduced by dimension based on metabonomics techniques. The true wine samples and false wine samples were correctly classified by clustering analysis. At the same time, the classification results of other wine series were basically consistent with the actual wine sample information. In the principal component analysis, PC1 and PC2 were extracted, and the variance information of 58.7% of the characteristic variables of different samples was explained, PC1 was 44.7%. The distinction between the real and the alcohol Hotelin ellipse regions was obvious and the visualization effect was intuitive. Other series of wine samples also keep the match of the real wine sample information. Partial least squares discriminant analysis was used to establish wine-related models. Samples of base wine were removed to highlight the material differences between real and fake wine. The table of importance of variables was obtained and the database of relevant substances was found. The first three most significant differences The characterizing compounds are butyl acetate, isoamyl hexanoate and 1-methylethylhexanoate.