Untargeted metabolite discovery in kinetic data from multi-dose intervention studies.

Area of Science:

• Metabolomics
• Biomarker Discovery
• Analytical Chemistry

Background:

• Metabolomics studies generate vast datasets, necessitating efficient methods for identifying relevant biomarkers.
• Distinguishing true biological signals from noise is challenging in complex biological samples.

Purpose of the Study:

• To present a new strategy for biomarker discovery using multi-dose kinetic metabolomics data.
• To develop a method for reducing large metabolomics datasets to a manageable list of relevant compounds.

Main Methods:

• Utilized gas chromatography-mass spectrometry (GC-MS) data in full scan mode.
• Analyzed time and dosage trends of compounds using principal component analysis.
• Focused on compounds exhibiting consistent trends across all doses and time points.

Main Results:

• Successfully reduced over 25,000 features to less than 250 in a gut fermentation study.
• Identified relevant metabolites by filtering for compounds with expected smooth time profiles across doses.
• Demonstrated the method's ability to flag relevant metabolites even with unique trends.

Conclusions:

• The proposed strategy effectively reduces large metabolomics datasets for biomarker discovery.
• This approach streamlines the identification of biologically meaningful metabolites.
• The method offers a significant reduction in the manual examination of complex data.