Metabolite annotation from knowns to unknowns through knowledge-guided multi-layer metabolic networking.

Area of Science:

• Metabolomics
• Bioinformatics
• Analytical Chemistry

Background:

• Untargeted metabolomics using liquid chromatography-mass spectrometry (LC-MS) is powerful for profiling metabolites.
• Annotation of unknown metabolites presents a significant challenge in untargeted metabolomics studies.
• Existing methods struggle with comprehensive identification of novel or uncharacterized compounds.

Purpose of the Study:

• To develop and validate a novel computational approach for global metabolite annotation, specifically addressing unknown compounds.
• To integrate diverse data types for enhanced accuracy in metabolite identification.
• To facilitate the discovery of the 'metabolomic dark matter'.

Main Methods:

• Development of a knowledge-guided multi-layer network (KGMN) integrating metabolic reaction, MS/MS similarity, and peak correlation networks.
• Application of KGMN to in vitro enzymatic reaction systems and diverse biological samples.
• Validation of annotated unknown metabolites using in silico MS/MS tools, repository mining, and chemical standard synthesis.

Main Results:

• KGMN successfully annotated approximately 100-300 putative unknown metabolites per dataset.
• >80% of these unknown metabolites were corroborated using in silico MS/MS analysis.
• Five previously unrepresented metabolites were validated through external data mining and synthesis.

Conclusions:

• The KGMN approach provides an efficient and effective strategy for annotating unknown metabolites in untargeted metabolomics.
• This method substantially advances the discovery of recurrent unknown metabolites in common biological samples.
• KGMN contributes to deciphering the vast unknown components within the metabolome.