基于熵最小算法的代谢产物GC-MS重叠色谱峰辨识  

Identification of components inside spectral overlapping peaks in metabolites GC-MS spectra by entropy minimization algorithms

作　　者：韦廷宗[1] 杜芳黎[1] 张华俊[1,2] 卢波[1] 郭媛[1] 林丽华[1] 张小平[2] 

机构地区：[1]广西工业生物技术研究中心,广西科学院,广西南宁530007 [2]苏州云溪分析技术有限公司,苏州工业园区,新加坡国立大学苏州研究院,江苏苏州215123

出　　处：《计算机与应用化学》2016年第5期587-592,共6页Computers and Applied Chemistry

基　　金：广西科技合作与交流计划资助项目(桂科合1347004-1)

摘　　要：代谢物的定性定量分析,是进一步探讨不同生理病理状态下的代谢物差异变化的基础,也是寻找生物标记物及其代谢通路的前提条件。但目前在对代谢产物GC-MS数据的分析中,重叠峰的解析却成为其分析的难题,重叠峰的存在使得色谱图无法直接进行定性和定量分析。对于重叠色谱峰,采用一般的色谱分析处理方法,大都耗时耗力,而随着计算技术的发展,各种新的方法不断被应用到代谢产物重叠峰的处理中。本文采用熵最小算法,对丙酮丁醇梭菌胞内代谢产物经硅烷化衍生后的GC-MS数据进行分析,对其中的几组重叠色谱峰进行重建分析,得到其所包含的纯组分,并对各组分分别进行定性和定量分析。结果表明,针对代谢产物复杂体系中的重叠色谱峰,采用熵最小方法能够比较全面的得到重叠信号中的各个组分的信息,并能成功地对其进行定性定量处理。Quantitative and qualitative analyses components inside metabolites samples, is a key prerequisite for finding metabolites differences under different physiological and pathological conditions, for locating biomarkers and finding metabolic routings as well. Now in metabolites GC-MS analyses, it is a headache to identify the components inside overlapped spectral peaks, qualitative and quantitative analyses are difficult as well. By current chemical analyses technologies, intensive labor works and great time have to be spend on dealing with those overlapped＇ peaks. With the development of computer technology, various new approaches have been applied to analyze the overlapped spectra in GC-MS analyses. In this work, entropy minimization algorithms have been used to reconstruct components out of overlapped peaks on silylanized metabolites samples. The results show that entropy minimization algorithms can do quantitative and qualitative analyses on components inside overlapped spectra peaks.

关 键 词：代谢产物 重叠峰 熵最小算法 定性分析 定量分析 

分 类 号：TQ015.9[化学工程] O6-39[理学—化学]