iModMix: integrative module analysis for multi-omics data

Area of Science:

• Computational Biology
• Bioinformatics
• Systems Biology

Background:

• Multi-omics data integration is crucial for understanding complex biological systems.
• Existing tools often rely on pathway annotations, limiting their scope and ability to handle unidentified features.
• There is a need for a flexible framework that can integrate diverse quantitative abundance data without prior biological knowledge.

Purpose of the Study:

• To develop a biology-agnostic framework for discovering novel associations across multiple omics datasets.
• To enable the integration of various quantitative abundance data types, including transcriptomics, proteomics, and metabolomics.
• To provide a user-friendly tool for both programming and non-programming users.

Main Methods:

• Constructs data-driven modules using graphical lasso to estimate sparse networks from omics features.
• Summarizes modules into eigenfeatures for horizontal integration across datasets.
• Correlates eigenfeatures across datasets while preserving the interpretability of individual omics types.

Main Results:

• iModMix enables the discovery of novel associations across transcriptomics, proteomics, and metabolomics data.
• The framework can seamlessly incorporate both identified and unidentified metabolites, overcoming limitations of existing metabolomics tools.
• Demonstrates utility in identifying novel multi-omics relationships in diverse biological contexts using public and in-house datasets.

Conclusions:

• iModMix is a versatile, biology-agnostic framework for multi-omics data integration.
• It facilitates the discovery of novel biological associations by leveraging data-driven modules.
• Available as an R package and a user-friendly R Shiny application, promoting accessibility for researchers.