Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada.
Anal Chem. 2013 Mar 19;85(6):3401-8. doi: 10.1021/ac400099b. Epub 2013 Mar 6.
Identification of unknown metabolites is a major challenge in metabolomics. Without the identities of the metabolites, the metabolome data generated from a biological sample cannot be readily linked with the proteomic and genomic information for studies in systems biology and medicine. We have developed a web-based metabolite identification tool ( http://www.mycompoundid.org ) that allows searching and interpreting mass spectrometry (MS) data against a newly constructed metabolome library composed of 8,021 known human endogenous metabolites and their predicted metabolic products (375,809 compounds from one metabolic reaction and 10,583,901 from two reactions). As an example, in the analysis of a simple extract of human urine or plasma and the whole human urine by liquid chromatography-mass spectrometry and MS/MS, we are able to identify at least two times more metabolites in these samples than by using a standard human metabolome library. In addition, it is shown that the evidence-based metabolome library (EML) provides a much superior performance in identifying putative metabolites from a human urine sample, compared to the use of the ChemPub and KEGG libraries.
鉴定未知代谢物是代谢组学面临的主要挑战。如果不知道代谢物的身份，那么从生物样本中生成的代谢组数据就无法轻易与蛋白质组和基因组信息联系起来，无法用于系统生物学和医学研究。我们开发了一个基于网络的代谢物鉴定工具（http://www.mycompoundid.org），它可以根据新构建的代谢物库，对质谱（MS）数据进行搜索和解释，该库由 8021 种已知的人类内源性代谢物及其预测代谢产物（来自一个代谢反应的 375809 种化合物和来自两个反应的 10583901 种化合物）组成。例如，在对人类尿液或血浆的简单提取物以及整个尿液进行液相色谱-质谱和 MS/MS 分析时，我们能够鉴定出这些样本中的代谢物，比使用标准人类代谢物库多至少两倍。此外，与使用 ChemPub 和 KEGG 库相比，基于证据的代谢物库（EML）在鉴定来自人类尿液样本的假定代谢物方面表现出更好的性能。
