MSACL 2024 Abstract

Propagation of Chemical Families from High-Confidence Level Metabolite Identification through Molecular Networking in the Context of Microbiome Research Romina PACHECO TAPIA (1, 2), Francesc PUIG CASTELLVI (1, 2), Inés CASTRO (1), Amélie Bonnefond (1), Philippe FROGUEL (1, 4), Marc-Emmanuel DUMAS (1, 2, 3, 5) (1) INSERM U1283, CNRS UMR 8199, Institut Pasteur de Lille, University of Lille, Lille University Hospital, Lille, France (2) Section of Biomolecular Medicine, Division of Systems Medicine, Department of Metabolism, Digestion and Reproduction, Imperial College London, London, UK (3) Section of Genomic & Environmental Medicine, National Heart & Lung Institute, Imperial College London (4) Department of Metabolism, Section of Genetics and Genomics, Imperial College London, London, UK (5) McGill Genome Centre, McGill University, Montréal, Qc, Canada. Romina Pacheco, Ph.D. (Presenter) Université de Lille >> POSTER (PDF)

Presenter Bio: Due to my interest in microbiology and biochemistry of our nature, I have worked with different types of microorganisms studying their pathogenicity but also their bioactive natural products. Thanks to the analytical tools I have discovered over time, I gained interest in metabolomics. During my PhD at the University of Toulouse, I applied a metabolomics approach using UHPLC-HRMS/MS to study the natural products of an endophytic fungal strain. During my current postdoc at the University of Lille, integrated into the metabolomics platform at EGID, I am highly motivated to apply and share my current knowledge in metabolomics in order to study the microbial and human metabolome and its link with health.

INTRODUCTION
Host-associated samples subjected to untargeted metabolomics have provided valuable insights into how microbes influence health in a bidirectional way[1,2]. However, accurate metabolite annotation and identification remain to be a challenge [3] along with ensuring analytical reproducibility and feature coverage for large cohorts of data [4].

OBJECTIVES
This study has the objective to improve the confidence level of the metabolite annotation by using chemical pure standards and in silico prediction tools in order to propagate their chemical classes to the unannotated metabolites.

METHODS
Human serum samples were analyzed using a UHPLC coupled to a high resolution Orbitrap Exploris 240 mass spectrometer with two optimized methods for polar and non-polar metabolites in negative and positive electrospray ionization mode. Intelligent data acquisition workflow was implemented in addition to the Data-Dependent Acquisition method in order to increase spectral data required for metabolite annotation. Ion Identity Molecular networking (IIMN) approach was applied using the GNPS on-line platform to expand the chemical class starting from the known metabolites, annotated with both public spectral reference library (GNPS) and an in-house spectral library, to the unknowns.

RESULTS
The metabolome profiling of the samples provided a total of 4840 linear and reproducible features (m/z-rt pairs) detected in positive and negative mode with both LC-MS methods. Focusing on the 410 unique metabolites that were annotated, 20% correspond to a high confidence level annotation. Among them, we identified lipids (45%), organic acids and derivatives (27%), organoheterocyclic compounds (10%), benzenoids (9%), organic oxygen compounds (5%) and other chemical superclasses, 38 classes and 76 subclasses. Propagation and in silico fragmentation tools allowed an increase of 20% of chemical categories assignment.

CONCLUSIONS
IIMN allowed us to reduce redundancies of ion species and to expand the chemical information of the unannotated metabolites. Metabolome mining tools, such as in silico approaches5, harness advanced machine learning and predict fragmentation spectra from known structures to complement our results. This will be essential for the implementation of a reproducible workflow for untargeted LCMS analysis of biofluids in the context of metabolomics in microbiome research. It will also help to increase and improve the identification of metabolites of interest to provide an appropriate biological interpretation.
References
1. Bauermeister, A., Mannochio-Russo, H., Costa-Lotufo, L.V., Jarmusch, A.K., Dorrestein, P.C., 2022. Nat. Rev. Microbiol. 20, 143–160.
2. Dekkers, K.F., Sayols-Baixeras, S., Baldanzi, G. et al., 2022. Nat Commun 13, 5370.
3. Plumb, R.S., Gethings, L.A., Rainville, P.D., Isaac, G., Trengove, R., King, A.M., Wilson, I.D., 2023. TrAC Trends Anal. Chem. 160, 116954.
4. Zhou, Z., Luo, M., Zhang, H., Yin, Y., Cai, Y., Zhu, Z.-J., 2022. Nat. Commun. 13, 6656.
5. Ernst, M.; Kang, K.B.; Caraballo-Rodríguez, A.M.; Nothias, L.-F.; Wandy, J.; Chen, C.; Wang, M.; Rogers, S.; Medema, M.H.; Dorrestein, P.C.; et al. 2019. Metabolites. 9, 144.