MS2Snoop: an R package for automatic extraction and curation of MS/MS spectra in metabolomics - ToxAlim Accéder directement au contenu
Poster De Conférence Année : 2022

MS2Snoop: an R package for automatic extraction and curation of MS/MS spectra in metabolomics

MS2Snoop: un package R pour automatiser l'extraction et le nettoyage des spectres MS/MS en métabolomique

Résumé

Introduction Omics approaches including metabolomics generate huge datasets requiring powerful processing tools. In the field of non-targeted mass spectrometry metabolomics, HRMS data mining has been the subject of much work, with the development of powerful interactive tools, including Workflow4Metabolomics (W4M). One of the main bottlenecks of current workflows lies in the complexity of the interpretation of results, in particular a lack of reliable MS/MS spectra reference databases to be used for confident identification (level 1 and 2 according to the metabolomics standard initiative). Creating reliable and well curated DB remains a laborious and challenging task, but necessary to ensure the quality of the reference spectra. In this work, we present a tool inspired by manual curation allowing the export of MS/MS data from targeted and data dependent (DDA) fragmentation methods to laboratory spectral databases, named “MS2Snoop”. Method Already integrated in W4M pipelines, MS2Snoop works with R and uses as starting point the peak list extracted from LC-MS and MS/MS analyses with XCMS 3.0 using W4M. Then, lists of precursors and fragment ions are processed with MSPurity using W4M, to assign fragmentation spectra to XCMS features, and to filter data on the purity of precursor ions. From these data, MS2Snoop has been developed to allow: • Retrieval of precursor and fragment ions data (m/z, RT, scan-by-scan intensities). • Matching of precursor ions with a list of expected compounds. • Establishing relationships between a precursor and its potential fragment ions via grouping, to obtain a raw spectrum. • Selecting a reference ion in the spectrum (precursor ion if still present, otherwise the most abundant ion). • Determining correlation between the reference ion and the other ions of the fragmentation spectrum using all scans. • Constructing a curated spectrum by eliminating ions with a correlation value lower than a threshold defined by the user. Results MS2Snoop was assessed on targeted MSMS analyses for about 30 compounds (pesticides and some of their metabolites) by performing the steps presented above. To validate these spectra, a manual curation was performed on the same data to compare the m/z and relative abundances of raw and MS2snoop treated spectra. The majority of the automatically extracted spectra using MS2Snoop showed similar results as spectra curated by an expert (fragment ions list, relative abundances), showing that MS2Snoop deconvolution process was efficient to provide automatically cleaned MS/MS spectra without interfering ions on all tested compounds. Conclusion MS2Snoop already allows extracting and cleaning MS/MS datasets using common data processing steps for metabolomics studies as a starting point, and then combined with inputs of a list of expected compounds and MS/MS raw data deconvoluted by MSPurity. Developed and validated with targeted MS/MS spectra, MS2Snoop is compatible with DDA spectra in the limit of the number of MS/MS scans acquired. This package will allow high throughput extraction of MS/MS spectra in order to feed the PeakForest project. Preliminary results obtained with MS2Snoop allowed considering it as a valuable tool for confident annotation of metabolites on chromatographic methods coupled to non-targeted mass spectrometry. A beta version is already publicly available on W4M.
Fichier principal
Vignette du fichier
Poster_MS2snoop_Analytics2022_vf.pdf (1.26 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
licence : CC BY NC SA - Paternité - Pas d'utilisation commerciale - Partage selon les Conditions Initiales

Dates et versions

hal-04001175 , version 1 (22-02-2023)

Licence

Paternité - Pas d'utilisation commerciale - Partage selon les Conditions Initiales

Identifiants

  • HAL Id : hal-04001175 , version 1

Citer

Kevin Wagner, Lain Pavot, Franck Giacomoni, Guillaume Marti, Yann Guitton, et al.. MS2Snoop: an R package for automatic extraction and curation of MS/MS spectra in metabolomics. Analytics 2022, Sep 2022, Nantes, France. 2022. ⟨hal-04001175⟩
56 Consultations
17 Téléchargements

Partager

Gmail Facebook X LinkedIn More