Abstract
Life’s chemical diversity far exceeds current biochemical maps. While metabolomics has catalogued tens of thousands of small molecules, conjugated metabolites, formed when two or more molecular entities are covalently fused through amidation, esterification, or related chemistries, remain underexplored. These molecules can act as microbial signals, detoxification intermediates, or endogenous regulators. Here, we mined 1.32 billion MS/MS spectra across public metabolomics repositories using reverse spectral searching coupled with delta-mass inference to map conjugation events. We generated structural hypotheses for 24,227,439 MS/MS clusters. From these, we inferred 217,291 substructure pairs with dual spectral support and 3,412,720 candidate conjugates with single-match support. Predictions span host–microbe co-metabolites, diet-derived conjugates, and drug-derived species, including drug-ethanolamine and creatinine conjugates with altered bioactivities. We also uncover a family of steroid-phosphoethanolamine conjugates. Fifty-five conjugates were matched by MS/MS of synthetic standards for this work, with 27 additionally supported by retention time matching in biological samples. Guidance on how to leverage this resource is also provided. Together, these results deliver a pan-repository map of potential conjugation chemistry, establish a resource for structural discovery and MS/MS annotation, and offer a scalable framework to explore the scope and diversity of the conjugated metabolome.
Full text
3,037 characters
· extracted from
oa-html
· click to expand
Abstract
Life’s chemical diversity far exceeds current biochemical maps. While metabolomics has catalogued tens of thousands of small molecules, conjugated metabolites, formed when two or more molecular entities are covalently fused through amidation, esterification, or related chemistries, remain underexplored. These molecules can act as microbial signals, detoxification intermediates, or endogenous regulators. Here, we mined 1.32 billion MS/MS spectra across public metabolomics repositories using reverse spectral searching coupled with delta-mass inference to map conjugation events. We generated structural hypotheses for 24,227,439 MS/MS clusters. From these, we inferred 217,291 substructure pairs with dual spectral support and 3,412,720 candidate conjugates with single-match support. Predictions span host–microbe co-metabolites, diet-derived conjugates, and drug-derived species, including drug-ethanolamine and creatinine conjugates with altered bioactivities. We also uncover a family of steroid-phosphoethanolamine conjugates. Fifty-five conjugates were matched by MS/MS of synthetic standards for this work, with 27 additionally supported by retention time matching in biological samples. Guidance on how to leverage this resource is also provided. Together, these results deliver a pan-repository map of potential conjugation chemistry, establish a resource for structural discovery and MS/MS annotation, and offer a scalable framework to explore the scope and diversity of the conjugated metabolome.
Competing Interest Statement
PCD is an advisor and holds equity in Cybele, BileOmix and Sirenas and a Scientific co-founder, advisor, holds equity and receives income to Ometa, Enveda, and Arome with prior approval by UC-San Diego. PCD also consulted for DSM animal health in 2023. LAB consulted for Locus Biosciences with prior approval from UC San Diego. MW is a co-founder of Ometa Labs LLC.
Data availability
All LC-MS/MS data involved in this study are publicly available on GNPS/MassIVE: mammalian feces (MSV000086131), diabetes study (MSV000082261), tomato seedling extracts (MSV000083306), microbial co-culturing (MSV000098638), human feces for losartan acylates (MSV000082433), human urine cohort for drug conjugates (MSV000096359), retention time matching between synthetic chemical standards and biological samples (MSV000099690).
Raw conjugate search results from 149.9 million clustered MS/MS spectra are available on Zenodo (https://zenodo.org/records/17245769). A rigorously curated subset was compiled into a reference MS/MS library and is accessible on GNPS2 (https://external.gnps2.org/gnpslibrary, “GNPS-CONJUGATED-METABOLOME”, four partitions). An interactive web application for exploring the conjugated metabolome is available at https://conjugated-metabolome.gnps2.org (also hosted at https://conjugated-metabolome.streamlit.app). The classical molecular networking job for tomatidine and its predicted conjugates is available at https://gnps2.org/status?task=06f0b0e04b1e4fe6b05deeb0de73ea7b.
Text is read by the "Ask this paper" AI Q&A widget below.
Extraction quality varies by source — PMC NXML preserves structure
cleanly, OA-HTML may include some navigation residue, and OA-PDF can
have broken hyphenation. The publisher copy
(via DOI)
is the canonical version.