CorrelationCalculator and Filigree: Tools for Data-Driven Network Analysis of Metabolomics Data

Area of Science:

• Biochemistry
• Bioinformatics
• Systems Biology

Background:

• Omics data analysis, particularly metabolomics, faces challenges in extracting actionable biological knowledge.
• Untargeted liquid chromatography-mass spectrometry (LC-MS) studies often contain numerous unknown metabolites, complicating the link between metabolite levels and biological processes.
• Existing pathway databases inadequately represent secondary and lipid metabolism.

Purpose of the Study:

• To present novel computational tools for data-driven network construction and analysis in metabolomics.
• To address limitations in current metabolomics data analysis, including unknown metabolites and incomplete pathway databases.
• To demonstrate the utility of CorrelationCalculator and Filigree for analyzing real-life metabolomics data.

Main Methods:

• Development of CorrelationCalculator for constructing single partial correlation-based networks from metabolomics data, even when metabolites exceed samples.
• Development of Filigree for building differential networks from two sample groups, followed by network clustering and enrichment analysis.
• Application of both tools to real-life metabolomics datasets to showcase their capabilities.

Main Results:

• CorrelationCalculator enables network construction from metabolomics data where samples are fewer than metabolites.
• Filigree facilitates differential network analysis, clustering, and enrichment, providing deeper biological insights.
• Both tools demonstrate practical utility in analyzing complex, real-world metabolomics data.

Conclusions:

• CorrelationCalculator and Filigree offer effective solutions for overcoming key challenges in metabolomics data analysis.
• These tools enhance the extraction of biological knowledge from omics data by enabling robust network construction and analysis.
• The presented methods improve the interpretation of metabolomics studies, particularly those involving unknown metabolites and underrepresented metabolic pathways.