【摘要】 代谢组学旨在对代谢组进行全景的分析,从而发现生物现象或规律.由于代谢物的种类和数量都非常庞大,对代谢组的分析高度依赖于分析仪器和方法.质谱是代谢组分析最有用的工具,超高效液相色谱-高分辨质谱可以从生物样品中获得大量的代谢物离子特征,获得丰富的代谢组信息,但如何对质谱数据进行挖掘和利用仍面临极大的挑战.计算代谢组学可以充分利用质谱采集的数据,结合统计、化学计量学、人工智能等方法实现对代谢组学数据的高效处理和分析,推动代谢组学的发展.本文在给出代谢组数据特点的基础上,综述了数据驱动的计算代谢组学方法学进展,包括特征提取、代谢物的注释和鉴定,并简要介绍了知识辅助的计算代谢组学方法,最后对计算代谢组学方法学下一步的发展进行了展望.更多还原

【Abstract】 Metabolomics aims at comprehensive analysis of the metabolome to discover biological phenomenon and mechanisms. Due to the complexity of metabolites in the biological samples, it is necessary to detect the metabolites as many as possible with advanced analytical techniques. Mass spectrometry is the most popular instrument in metabolomics research. The abundant ion signals from mass spectrometry are reflections of metabolome. However, it is still a huge challenge to process data and mine information from mass spectrometry data. Computational metabolomics fully exploits data from instruments, then combines methods from statics, chemometrics or artificial intelligence to make data processing and analysis efficiently, and therefore promotes the development of metabolomics. In this review, we firstly introduce general metabolomics data analysis procedure on the basis of the characteristic of data. Then datadriven computational metabolomics methods are reviewed, including feature detection, metabolites identification and annotation. Finally, knowledge-assisted computational metabolomics methods and future perspectives are briefly presented.更多还原