From MS1 to Structure: A Van Krevelen–DBE–Aromaticity-Based Framework for Annotating Specialized Metabolites via High-Resolution Mass Spectrometry

Abstract

Rationale

Classifying specialized metabolites in untargeted metabolomics remains a major challenge, particularly when relying solely on high-resolution mass spectrometry (HRMS) data at the MS1 level. Traditional approaches using Van Krevelen diagrams often lack sufficient resolution to distinguish structurally similar metabolite classes.

Methods

We developed a chemoinformatic framework that combines Van Krevelen analysis (H/C vs. O/C) with double bond equivalent (DBE) calculations to refine metabolite class annotation at Level 3 of the Metabolomics Standards Initiative (MSI). Molecular formulas were retrieved from curated structure databases and natural product repositories, and DBE values were used to refine structural classification. A dataset of over 600 curated molecular formulas representing phenolics, alkaloids, and isoprenoids was analyzed to define class-specific patterns.

Results

The combined use of DBE and Van Krevelen plots enabled improved discrimination between overlapping metabolite classes, including flavonoids, phenolic acids, coumarins, and tannins. Our framework revealed structural trends associated with aromaticity and unsaturation that are not captured by conventional MS1-based tools. It outperforms existing Level 3 annotation strategies that rely on in silico MS/MS fragmentation or substructure matching. A case study using Eugenia jambolana fruit extract validated the method, revealing dominant classes such as flavonoids, phenolic acids, and tannins using only MS1 data.

Conclusions

This is the first scalable framework to annotate specialized metabolites from MS1 data alone using integrated elemental ratios and structural descriptors. It enhances the annotation confidence for untargeted metabolomics, especially in complex, undercharacterized plant matrices, without requiring MS2 fragmentation.