DIANMF: Combined MS and MS/MS deconvolution of SWATH DIA data with the DIA-NMF workflow for comprehensive annotation in metabolomics
DIANMF is an open-source R package for deconvolving complex SWATH-DIA metabolomics data using sparse non-negative matrix factorization (NMF). It is the first method to jointly unmix MS1 and MS2 spectra in a fully untargeted manner, without relying predefined peak models or on spectral libraries. DIA-NMF enables precursor-level interpretation by recovering pure MS1 spectra and enriched, unmixed MS2 fragmentation patterns from all relavent isolation windows. This improves compound identification, especially for co-eluting and low-intensity metabolites.
The workflow detects MS1 peaks, extracts mixed MS1 and MS2 signals within minimally overlapping retention time windows, aligns them, and jointly unimixes these to recover pure precursor and fragment patterns (MS1 and MS2 spectra, respectively).
The package can be installed from GitHub with:
#install.packages("devtools")
devtools::install_github("odisce/DIANMF")The processing of SWATH-DIA metabolomics data with the DIANMF package is described in the included vignette (vignettes/Process_SWATH_DIA_Data.html).
This package includes a sample dataset used in the accompanying examples and vignettes. It consists of selected regions from replicates 1 and 2 of a 10 ng/mL spiked human plasma sample, focusing on the 280–320 m/z range and 280–320 seconds retention time window (Barbier Saint-Hilaire et al. 2020).
Diana Karaki, Annelaure Damon, Antoine Souloumiac, François Fenaille, Etienne A. Thévenot and Sylvain Dechaumet (in preparation). Combined MS and MS/MS deconvolution of SWATH DIA data with the DIA-NMF workflow for comprehensive annotation in metabolomics.
[email protected], [email protected], and [email protected]
Karaki et al. (2024) Non-Negative Matrix Factorization of SWATH DIA Data Improves Global Metabolite Identification. European Signal Processing Conference (EUSIPCO), 2387-2391. DOI:10.23919/EUSIPCO63174.2024.10715181.
Wang et al. (2019) Advancing untargeted metabolomics using data-independent acquisition mass spectrometry technology. Analytical and bioanalytical chemistry, 4349-4357. DOI:10.1007/s00216-019-01709-1.
Barbier Saint-Hilaire et al. (2020) Comparative Evaluation of Data Dependent and Data Independent Acquisition Workflows Implemented on an Orbitrap Fusion for Untargeted Metabolomics. Metabolites, 2218-1989. DOI:10.3390/metabo10040158.