Large-scale discovery and annotation of substructure patterns in mass spectrometry profiles
7 affiliations
- 1Bioinformatics Group, Wageningen University & Research, Wageningen, the Netherlands. [email protected]
- 2Bioinformatics Group, Wageningen University & Research, Wageningen, the Netherlands. [email protected]
- 3Wageningen Food Safety Research - part of Wageningen University & Research, Wageningen, the Netherlands. [email protected]
- 4College of Medical, Veterinary & Life Sciences, University of Glasgow, Glasgow, UK
- 5Wageningen Food Safety Research - part of Wageningen University & Research, Wageningen, the Netherlands
- 6Bioinformatics Group, Wageningen University & Research, Wageningen, the Netherlands. [email protected]
- 7Department of Biochemistry, University of Johannesburg, Johannesburg, South Africa. [email protected]
Abstract
Untargeted mass spectrometry can detect thousands of molecules at once, potentially offering powerful insights into complex samples. However, the increasing scale of experimental datasets and spectral libraries limits our ability to extract and annotate structural information to allow for interpretation. Here, we present the software tool MS2LDA 2.0 that helps to address this gap by identifying recurring fragmentation patterns (Mass2Motifs) that can reflect shared chemical substructures. We introduce automated annotation support through Mass2Motif Annotation Guidance (MAG) that provides suggestions to interpret detected patterns. Our unsupervised pattern mining tool enables the study of much larger datasets with up to 14 times faster analysis than its predecessor. We demonstrate the utility of MS2LDA 2.0 and MAG in applications such as detecting pesticide-related substructures and exploring unknown fungal compounds. Together, these advances make it easier to uncover meaningful chemical patterns in complex data.
Related
- scBOA: Bayesian optimization of single-cell analysis pipelines for robust annotation and discovery
- scGlycoAI: machine learning analysis of single-cell glycan and RNA sequencing data for cell-type annotation and malignancy-associated glycan discovery
- Artificial Intelligence in Mass Spectrometry-Based Metabolomics and Lipidomics for Drug Discovery and Development
- Discovery and annotation of putative genes and novel RNA splice-isoforms in the human heart.
- Dietary patterns in early pregnancy are associated with distinct gut and vaginal microbial profiles: Insights from the MicrobeMom study
BibTeX
@article{ortega2026largescale,
title = {{Large-scale discovery and annotation of substructure patterns in mass spectrometry profiles}},
author = {Laura Rosina Torres Ortega and Jonas Dietrich and Joe Wandy and Hans G.J. Mol and Justin J. J. van der Hooft},
journal = {Nature Communications},
year = {2026},
month = jul,
doi = {10.1038/s41467-026-75038-0},
url = {https://doi.org/10.1038/s41467-026-75038-0},
note = {CC BY},
}
RIS
TY - JOUR
TI - Large-scale discovery and annotation of substructure patterns in mass spectrometry profiles
AU - Laura Rosina Torres Ortega
AU - Jonas Dietrich
AU - Joe Wandy
AU - Hans G.J. Mol
AU - Justin J. J. van der Hooft
PY - 2026
DA - 2026/07/06
JO - Nature Communications
DO - 10.1038/s41467-026-75038-0
UR - https://doi.org/10.1038/s41467-026-75038-0
AB - Untargeted mass spectrometry can detect thousands of molecules at once, potentially offering powerful insights into complex samples. However, the increasing scale of experimental datasets and spectral libraries limits our ability to extract and annotate structural information to allow for interpretation. Here, we present the software tool MS2LDA 2.0 that helps to address this gap by identifying recurring fragmentation patterns (Mass2Motifs) that can reflect shared chemical substructures. We introduce automated annotation support through Mass2Motif Annotation Guidance (MAG) that provides suggestions to interpret detected patterns. Our unsupervised pattern mining tool enables the study of much larger datasets with up to 14 times faster analysis than its predecessor. We demonstrate the utility of MS2LDA 2.0 and MAG in applications such as detecting pesticide-related substructures and exploring unknown fungal compounds. Together, these advances make it easier to uncover meaningful chemical patterns in complex data.
SN - 2041-1723
ER -