Automated annotation and evaluation of in-source mass spectra in GC/atmospheric pressure chemical ionization-MS-based metabolomics

Tools

Item Type:	Article
Title:	Automated annotation and evaluation of in-source mass spectra in GC/atmospheric pressure chemical ionization-MS-based metabolomics
Creators Name:	Jaeger, C., Hoffmann, F., Schmitt, C.A. and Lisec, J.
Abstract:	Gas chromatography using atmospheric pressure chemical ionization coupled to mass spectrometry (GC/APCI-MS) is an emerging metabolomics platform, providing much-enhanced capabilities for structural mass spectrometry as compared to traditional electron ionization (EI)-based techniques. To exploit the potential of GC/APCI-MS for more comprehensive metabolite annotation, a major bottleneck in metabolomics, we here present the novel R-based tool InterpretMSSpectrum assisting in the common task of annotating and evaluating in-source mass spectra as obtained from typical full-scan experiments. After passing a list of mass-intensity pairs, InterpretMSSpectrum locates the molecular ion (M0), fragment, and adduct peaks, calculates their most likely sum formula combination, and graphically summarizes results as an annotated mass spectrum. Using (modifiable) filter rules for the commonly used methoximated-trimethylsilylated (MeOx-TMS) derivatives, covering elemental composition, typical substructures, neutral losses, and adducts, InterpretMSSpectrum significantly reduces the number of sum formula candidates, minimizing manual effort for postprocessing candidate lists. We demonstrate the utility of InterpretMSSpectrum for 86 in-source spectra of derivatized standard compounds, in which rank-1 sum formula assignments were achieved in 84% of the cases, compared to only 63% when using mass and isotope information on the M0 alone. We further use, for the first time, automated annotation to evaluate the purity of pseudospectra generated by different metabolomics preprocessing tools, showing that automated annotation can serve as an integrative quality measure for peak picking/deconvolution methods. As an R package, InterpretMSSpectrum integrates flexibly into existing metabolomics pipelines and is freely available from CRAN (https://cran.r-project.org/).
Keywords:	Algorithms, Gas Chromatography-Mass Spectrometry, Metabolomics, Molecular Weight, Software
Source:	Analytical Chemistry
ISSN:	0003-2700
Publisher:	American Chemical Society
Volume:	88
Number:	19
Page Range:	9386-9390
Date:	4 October 2016
Official Publication:	https://doi.org/10.1021/acs.analchem.6b02743
PubMed:	View item in PubMed

Repository Staff Only: item control page