Research On New UPLC-HRMS Data Analysis Methods With Application In Glycyrrhiza Uralensis Fisch

Non-targeted metabolic profiling technology based on ultra performance liquid chromatography coupled with high-resolution mass spectrometry(UPLC-HRMS)has been extensively used in many scientific fields.With the continuous improvement of sensitivity,chromatographic resolution and mass spectrometry accuracy,researchers are able to collect more and more abundant compound information.However,how to accurately extract compounds from the data has become a bottleneck.The purpose of this thesis is to develop a noval UPLC-HRMS data analysis scheme,which is intended to solve the problems of false positives and false negatives existing in data analysis and the analysis of profile mode data.The research content of the subject research mainly includes the following three parts:(1)Constructing a new EIC construction strategy based on the combination of intensity and density and a chromatographic peak extraction strategy based on the combination of smoothing and window extending to implement accurate EIC construction and chromatographic peak extraction;(2)Based on MATLAB and C++ platforms,a toolkit is developed for UPLC-HRMS profile mode data analysis.It is used to perform analysis of profile mode data obtained by instruments;(3)Based on the Ant DASProfiler data analysis toolkit developed above,it is used to distinguish the origins of licorice from 4 different province,and to identify the differential metabolites among them.The main research results are as follows:Part Ⅰ: A new EIC construction strategy based on the combination of intensity and density and a new peak extraction strategy based on smoothing and window extending were constructed.It is implemented based on MATLAB and C ++ platform and embedded in the data analysis toolkit Ant DAS that we developed previously to perform data analysis.The analysis of tea,licorice and tobacco leaves from real plant samples demonstrates the superior performance of the new method.The distribution of ions in the actual example of EIC construction and the choice of the center point of the chromatographic peak can well prove the rationality of our proposed EIC construction strategy based on the combination of intensity and density.Compared with the other four existing EIC construction algorithms,the new EIC construction method can achieve more accurate EIC construction,which helps reduce the occurrence of false positives and false negatives.Compared with XCMS and Ant DAS,the new method of peak extraction can achieve more accurate extraction of chromatographic peaks and judge the position of chromatographic peaks.For resolving coeluting peaks,the new method can also accurately identify and reduce the probability of false positives and false negatives.Part Ⅱ: Ant DASProfiler software was developed using MATLAB and C ++ to perform complete UPLC-HRMS profile data analysis consists of several stages: EIC construction,peak extraction,peak alignment,annotation and fragment ion identification,and statistical analysis.By analyzing the tea and licorice data obtained from Agilent,Thermo and Waters UPLC-HRMS instruments in centroid and profile mode,it is found that the profile data of Agilent and Thermo can obtain richer compound information,while the compound information obtained by Waters’ centroid mode is richer,but the impact of noise is very large.This result will provide a reference for users to choose a data acquisition mode when using UPLC-HRMS instruments.Ant DASProfiler compares the XCMS analysis results.The compound information obtained in the XCMS analysis results is more abundant,but among the unique parts of XCMS with high correlation coefficients,there are almost noisy peaks.Part Ⅲ: UPLC-HRMS-based non-targeted metabolomics was used to analyze licorice samples from four different origins in Neimenggu,Ningxia,Xinjiang,and Gansu.The Ant DASProfiler software toolkit developed by us was used to complete the data analysis,and the differential metabolites of the compound ion list were screened in two ion modes.Among them,135 differential metabolites were screened in the positive ion mode and 107 in the negative ion mode.Selecting metabolites with rich fragment ion information and fuse the compound information in two ion modes to obtain:(1)18unique compounds in positive ion mode;(2)unique differences in negative ion mode 7compounds;(3)26 differential compounds in positive and negative ion mode.A Fisher discriminantion analysis was used to establish the origin identification model,and the licorice samples from four different licorice origin regions were distinguished.Xinjiang is most clearly distinguished due to its location in the most northwest.Compound library matching and literature search were used to identify compounds that co-existed in the two modes and compounds that existed in only one mode.Eight differential metabolites were identified in total.