mzGroupAnalyzer--predicting pathways and novel chemical structures from untargeted high-throughput metabolomics data

Area of Science:

• Metabolomics
• Systems Biology
• Plant Science

Background:

• The metabolome reflects genotype-environment-phenotype interactions, but identifying unknown metabolites is a major challenge.
• Manual analysis of large metabolomics datasets is infeasible, necessitating automated computational approaches.
• Understanding metabolite dynamics is crucial for deciphering biochemical transformations.

Purpose of the Study:

• To develop an automated pathway inference strategy for analyzing metabolomics time-series data.
• To enable automated m/z feature extraction and structure or pathway assignment.
• To identify novel compounds and visualize metabolic transformations.

Main Methods:

• Utilized liquid chromatography-mass spectrometry (LC-MS) with high resolution and mass accuracy for metabolome time-series measurements.
• Developed the mzGroupAnalyzer algorithm for automated detection of metabolite transformations.
• Employed van Krevelen diagrams to visualize oxidative processes and biochemical transformations.

Main Results:

• Applied the method to Arabidopsis thaliana under cold and high light stress, observing a shift to purple color due to flavonoid accumulation.
• Identified 15 putatively novel compounds involved in the flavonoid pathway.
• Validated new compounds using product ion spectra from the same dataset.

Conclusions:

• The automated strategy successfully identifies metabolite transformations and potential novel compounds in complex biological systems.
• The mzGroupAnalyzer, integrated into the COVAIN toolbox, offers a powerful approach for metabolomics data analysis.
• This method can be extended to diverse biological systems for pathway inference and discovery.