中文摘要：

采用化学计量学方法对气相色谱-质谱联用代谢指纹分析数据进行处理,通过微矩阵显著性分析策略寻找不同碳源条件下6株不同遗传背景大肠杆菌变化最显著代谢物,以所发现的最显著变化的代谢物为中心,分别与同条件下检出的其他代谢物进行相关分析,筛选到相关系数较大的代谢物。不仅发现了在不同培养条件下波动显著性化合物（葡萄糖和果糖为唯一碳源条件下变化最显著代谢物分别为谷氨酰胺和尸胺）,并通过聚类分析验证了该数据处理方法的有效性,保证了在大量数据中提取到更加有用的信息。

英文摘要：

As a branch of chemical science,chemometrics is particularly suitable for dealing with a large number of complex data produced by GC-MS.In this study,a method based on metabolic fingerprinting analysis was established.Chemometrics method was used to deal with GC-MS data.As a result,compounds that highly fluctuate in the different culture conditions were discovered,and the validity of the data processing method was proved by cluster analysis,which indicated that the method can be used to extract more useful information from the large amounts of data.