Assessment of data pre-processing methods for LC-MS/MS-based metabolomics of uterine cervix cancer
- 22 February 2013
- journal article
- research article
- Published by Royal Society of Chemistry (RSC) in The Analyst
- Vol. 138 (9) , 2669-2677
Abstract
A metabolomics strategy based on rapid resolution liquid chromatography/tandem mass spectrometry (RRLC-MS/MS) and multivariate statistics has been implemented to identify potential biomarkers in uterine cervix cancer. Due to the importance of the data pre-processing method, three popular software packages have been compared. Then they have been used to acquire respective data matrices from the same LC-MS/MS data. Multivariate statistics was subsequently used to identify significantly changed biomarkers for uterine cervix cancer from the resulting data matrices. The reliabilities of the identified discriminated metabolites have been further validated on the basis of manually extracted data and ROC curves. Nine potential biomarkers have been identified as having a close relationship with uterine cervix cancer. Considering these in combination as a biomarker group, the AUC amounted to 0.997, with a sensitivity of 92.9% and a specificity of 95.6%. The prediction accuracy was 96.6%. Among these potential biomarkers, the amounts of four purine derivatives were greatly decreased, which might be related to a P2 receptor that might lead to a decrease in cell number through apoptosis. Moreover, only two of them were identified simultaneously by all of the pre-processing tools. The results have demonstrated that the data pre-processing method could seriously bias the metabolomics results. Therefore, application of two or more data pre-processing methods would reveal a more comprehensive set of potential biomarkers in non-targeted metabolomics, before a further validation with LC-MS/MS based targeted metabolomics in MRM mode could be conducted.Keywords
This publication has 26 references indexed in Scilit:
- Time-Course Changes in Potential Biomarkers Detected Using a Metabonomic Approach in Walker 256 Tumor-Bearing RatsJournal of Proteome Research, 2011
- MZmine 2: Modular framework for processing, visualizing, and analyzing mass spectrometry-based molecular profile dataBMC Bioinformatics, 2010
- Integrated Ionization Approach for RRLC−MS/MS-based Metabonomics: Finding Potential Biomarkers for Lung CancerJournal of Proteome Research, 2010
- Metabonomics study of liver cancer based on ultra performance liquid chromatography coupled to mass spectrometry with HILIC and RPLC separationsAnalytica Chimica Acta, 2009
- Metabolomic profiles delineate potential role for sarcosine in prostate cancer progressionNature, 2009
- Preprocessing, classification modeling and feature selection using flow injection electrospray mass spectrometry metabolite fingerprint dataNature Protocols, 2008
- Visualization of GC/TOF-MS-Based Metabolomics Data for Identification of Biochemically Interesting Compounds Using OPLS Class ModelsAnalytical Chemistry, 2007
- The P2X7 Receptor: A Novel Biomarker of Uterine Epithelial CancersCancer Epidemiology, Biomarkers & Prevention, 2006
- XCMS: Processing Mass Spectrometry Data for Metabolite Profiling Using Nonlinear Peak Alignment, Matching, and IdentificationAnalytical Chemistry, 2006
- Acylcarnitines in intermediary metabolismEuropean Journal of Pediatrics, 1995