MetaboT: an LLM-based multi-agent framework for interactive analysis of mass spectrometry metabolomics knowledge graphs.
15 affiliations
- 1Université Côte d'Azur, CNRS, ICN, Nice, France.
- 2Interdisciplinary Institute for Artificial Intelligence (3iA) Côte d'Azur, Sophia-Antipolis, France.
- 3Department of Computer Science, University of Antwerp, Antwerp, Belgium.
- 4Massachusetts Institute of Technology, Cambridge, MA, USA.
- 5probabl., Paris, France.
- 6Inria, Université Côte d'Azur, CNRS, I3S, Nice, France.
- 7Department of Computer Science, Tufts University, Medford, MA, 02155, USA.
- 8Institute of Pharmaceutical Sciences of Western Switzerland, University of Geneva, Centre Médical Universitaire, Geneva, Switzerland.
- 9School of Pharmaceutical Sciences, University of Geneva, Centre Médical Universitaire, Geneva, Switzerland.
- 10Swiss Institute of Bioinformatics (SIB), Lausanne, Switzerland.
- 11Department of Chemical and Biological Engineering, Tufts University, Medford, MA, 02155, USA.
- 12Department of Computer Science, University of Antwerp, Antwerp, Belgium. [email protected].
- 13Inria, Université Côte d'Azur, CNRS, I3S, Nice, France. [email protected].
- 14Université Côte d'Azur, CNRS, ICN, Nice, France. [email protected].
- 15Interdisciplinary Institute for Artificial Intelligence (3iA) Côte d'Azur, Sophia-Antipolis, France. [email protected].
Abstract
Mass spectrometry-based metabolomics generates complex, high-dimensional data that holds vast potential for biological discovery but remains difficult to integrate and interpret. Knowledge graphs (KGs) unify this heterogeneous information by representing spectra, annotations, taxa, chemical classes, and biological activities as a single interoperable network; however, their practical use is limited by the steep learning curve of corresponding specialized representation and query languages. Here we introduce MetaboT, an open-source multi-agent Large Language Model (LLM) framework that translates natural-language questions into executable SPARQL queries over metabolomics knowledge graphs. MetaboT is designed to mitigate the hallucination and schema-compliance limitations of single-model approaches through a modular architecture in which specialised agents handle scope validation, entity resolution against authoritative resources, schema-aware query generation, iterative refinement, and result interpretation. We validated MetaboT on the Experimental Natural Products Knowledge Graph (ENPKG), using an expert-authored benchmark of natural-language questions paired with reference SPARQL queries, and demonstrate its ability to answer complex questions about plant-metabolite relationships and biological activities, achieving 83.67% accuracy versus 8.16% for a single-shot baseline and reducing schema-mismatch errors to a single residual case in our benchmark. MetaboT lowers the technical barrier for metabolomics researchers and enables semantic data mining without specialised programming expertise.Scientific contributionMetaboT introduces a multi-agent LLM architecture that grounds each step of natural-language-to-SPARQL translation in authoritative identifier resolution and schema-aware validation, substantially reducing the hallucinations observed with single-model baselines. The framework couples orchestrated agents with domain-specific tools for chemical, taxonomic, and biological target resolution, and with an iterative refinement loop that distinguishes query-construction errors from genuine data absence. Released as open-source software with an archived evaluated version, a reusable benchmark, and a public web demonstrator, MetaboT provides a reproducible reference implementation for knowledge-graph-driven exploration of mass spectrometry metabolomics data.
Related
- Truthful visualizations for mass spectrometry imaging enable high-spatial-resolution interactive m/z mapping and exploration
- Data processing in proteomics and metabolomics: liquid chromatography-mass spectrometry-based omics with divergent computational paradigms.
- Chainable DataFrame Framework for Interactive Chemical Data Processing with pdChemChain
- Integrated High-Throughput Mass Spectrometry Framework for Multiplexed Substrate Screening in Enzyme Discovery and Development
- Mass Spectrometry-Based Omics: An Integrated Framework for Analytical Quality, Data Integrity, and Biological Interpretation
BibTeX
@article{bekbergenova2026metabot,
title = {{MetaboT: an LLM-based multi-agent framework for interactive analysis of mass spectrometry metabolomics knowledge graphs.}},
author = {Madina Bekbergenova and Lucas Pradi and Benjamin Navet and Emma Tysinger and Franck Michel and Matthieu Feraud and Yousouf Taghzouti and Yan Zhou Chen and Olivier Kirchhoffer and Florence Mehl and Martin Legrand and Tao Jiang and Marco Pagni and Soha Hassoun and Jean-Luc Wolfender and Wout Bittremieux and Fabien Gandon and Louis-Félix Nothias},
journal = {Journal of cheminformatics},
year = {2026},
month = jul,
doi = {10.1186/s13321-026-01257-8},
url = {https://europepmc.org/article/MED/42907998},
note = {CC BY},
}
RIS
TY - JOUR
TI - MetaboT: an LLM-based multi-agent framework for interactive analysis of mass spectrometry metabolomics knowledge graphs.
AU - Madina Bekbergenova
AU - Lucas Pradi
AU - Benjamin Navet
AU - Emma Tysinger
AU - Franck Michel
AU - Matthieu Feraud
AU - Yousouf Taghzouti
AU - Yan Zhou Chen
AU - Olivier Kirchhoffer
AU - Florence Mehl
AU - Martin Legrand
AU - Tao Jiang
AU - Marco Pagni
AU - Soha Hassoun
AU - Jean-Luc Wolfender
AU - Wout Bittremieux
AU - Fabien Gandon
AU - Louis-Félix Nothias
PY - 2026
DA - 2026/07/01
JO - Journal of cheminformatics
DO - 10.1186/s13321-026-01257-8
UR - https://europepmc.org/article/MED/42907998
AB - Mass spectrometry-based metabolomics generates complex, high-dimensional data that holds vast potential for biological discovery but remains difficult to integrate and interpret. Knowledge graphs (KGs) unify this heterogeneous information by representing spectra, annotations, taxa, chemical classes, and biological activities as a single interoperable network; however, their practical use is limited by the steep learning curve of corresponding specialized representation and query languages. Here we introduce MetaboT, an open-source multi-agent Large Language Model (LLM) framework that translates natural-language questions into executable SPARQL queries over metabolomics knowledge graphs. MetaboT is designed to mitigate the hallucination and schema-compliance limitations of single-model approaches through a modular architecture in which specialised agents handle scope validation, entity resolution against authoritative resources, schema-aware query generation, iterative refinement, and result interpretation. We validated MetaboT on the Experimental Natural Products Knowledge Graph (ENPKG), using an expert-authored benchmark of natural-language questions paired with reference SPARQL queries, and demonstrate its ability to answer complex questions about plant-metabolite relationships and biological activities, achieving 83.67% accuracy versus 8.16% for a single-shot baseline and reducing schema-mismatch errors to a single residual case in our benchmark. MetaboT lowers the technical barrier for metabolomics researchers and enables semantic data mining without specialised programming expertise.Scientific contributionMetaboT introduces a multi-agent LLM architecture that grounds each step of natural-language-to-SPARQL translation in authoritative identifier resolution and schema-aware validation, substantially reducing the hallucinations observed with single-model baselines. The framework couples orchestrated agents with domain-specific tools for chemical, taxonomic, and biological target resolution, and with an iterative refinement loop that distinguishes query-construction errors from genuine data absence. Released as open-source software with an archived evaluated version, a reusable benchmark, and a public web demonstrator, MetaboT provides a reproducible reference implementation for knowledge-graph-driven exploration of mass spectrometry metabolomics data.
SN - 1758-2946
ER -