Enhancing untargeted metabolomics using metadata-based source annotation

Area of Science:

• Metabolomics
• Biochemistry
• Bioinformatics

Background:

• Untargeted metabolomics studies in humans typically identify only a small fraction (~10%) of detected molecular features.
• Accurate annotation of these features is crucial for understanding biological processes and dietary impacts.

Purpose of the Study:

• To introduce and validate a novel reference-data-driven analysis approach for human metabolomics.
• To enhance the utilization of tandem mass spectrometry (MS/MS) data for feature annotation.
• To enable empirical assessment of dietary patterns using untargeted metabolomics data.

Main Methods:

• Developed a method to match untargeted metabolomics MS/MS data against metadata-annotated source data, creating a pseudo-MS/MS reference library.
• Applied this approach to analyze food source data.
• Compared the performance against conventional structural MS/MS library matching.

Main Results:

• The reference-data-driven approach increased MS/MS spectral usage 5.1-fold compared to conventional library matching.
• Successfully enabled empirical assessment of dietary patterns from untargeted metabolomics data.
• Significantly improved the annotation rate of molecular features in human metabolomics studies.

Conclusions:

• Reference-data-driven analysis is a powerful strategy to improve feature annotation in human metabolomics.
• This method enhances the utility of untargeted metabolomics for investigating dietary intake and patterns.
• The approach offers a significant advancement over existing MS/MS library matching techniques.