Automated Optimization of XCMS Parameters for Improved Peak Picking of LC-MS Data using the Coefficient of Variation and Parameter Sweeping for Untargeted Metabolomic

View Researcher's Other Codes

Disclaimer: The provided code links for this paper are external links. Science Nest has no responsibility for the accuracy, legality or content of these links. Also, by downloading this code(s), you agree to comply with the terms of use as set out by the author(s) of the code(s).

Please contact us in case of a broken link from here

Authors Sascha K Manier, Andreas Keller, Markus R Meyer
Journal/Conference Name Drug Testing and Analysis
Paper Category
Paper Abstract Accurate peak picking and further processing is a current challenge in the analysis of untargeted metabolomics using liquid chromatography-mass spectrometry (LC-MS) data. The optimization of these processes is crucial to obtain proper results. This study investigated and optimized the detection of peaks by XCMS, a widely used R package for peak picking and processing of high-resolution LC-MS metabolomics data by their coefficient of variation using neat standard solutions of drug like compounds. The obtained results were additionally verified by using fortified pooled plasma samples. Settings of the mass spectrometer were optimized by recommendations in literature to enable a reliable detection of the investigated analytes. XCMS parameters were evaluated using a comprehensive parameter sweeping approach. The optimization steps were statistically evaluated and further visualized after principal component analysis (PCA). Concerning the lower concentrated solution in methanol samples, the optimization of both mass spectrometer and XCMS parameters improved the median coefficient of variation from 24% to 7%, retention time fluctuation from 9.3 seconds to 0.54 seconds, and fluctuation of the mass to charge ratio (m/z) from m/z 0.00095 to m/z 0.00028. The number of parent compounds and their related species annotated by CAMERA increased from 88 to 113 while the total amount of features decreased from 3282 to 428. Optimized MS settings such as increased resolution led to a higher specificity of peak picking. PCA supported these findings by showing the best clustering of samples after optimization of both mass spectrometer and XCMS parameters. The results implied that peak picking needs to be individually adapted for the experimental set up. Reducing unwanted variation in the data set was most successful after combining high resolving power with strict peak picking settings.
Date of publication 2018
Code Programming Language R

Copyright Researcher 2022