iModMix

iModMix enables integrative module analysis for multi-omics data by constructing data-driven networks and correlating modules across heterogeneous datasets without relying on prior biological knowledge or pathway annotations.

Key Features:

• Biology-agnostic framework: Discovers novel associations across any type of quantitative abundance data without relying on pathway annotations or prior biological knowledge.
• Graphical lasso network estimation: Constructs sparse networks from omics features using graphical lasso to create data-driven modules.
• Eigenfeature summarization: Summarizes modules into eigenfeatures for horizontal integration across datasets while preserving distinct feature sets and interpretability of each omics type.
• Unidentified metabolite incorporation: Seamlessly incorporates both identified and unidentified metabolites because it does not rely on annotations such as KEGG identifiers.
• Direct matrix operation: Operates directly on matrices containing expression or abundances for features including genes, proteins, and metabolites.

Scientific Applications:

• Multi-omics integration: Analyzes metabolomics, proteomics, and transcriptomics data to explore intricate molecular associations within heterogeneous biological systems.
• Novel association discovery: Identifies novel multi-omics relationships in diverse biological contexts across any type of quantitative abundance data.

Methodology:

iModMix constructs data-driven modules using graphical lasso to estimate sparse networks from omics features, summarizes these modules into eigenfeatures, and correlates them across datasets for horizontal integration.