MulNet: a scalable framework for reconstructing intra- and intercellular signaling networks from bulk and single-cell...

Area of Science:

• Bioinformatics
• Systems Biology
• Genomics

Background:

• Gene expression regulation is complex, involving multiple molecular interactions.
• Existing molecular networks often use limited interaction types, leading to incomplete understanding.
• A comprehensive approach is needed to model these diverse interactions for better gene regulation insights.

Purpose of the Study:

• To introduce MulNet, a novel framework for constructing multilayer networks from diverse molecular interactions.
• To enable accurate identification of gene modules and key regulators within these networks.
• To apply MulNet to cancer datasets for identifying novel therapeutic targets and understanding cellular communication.

Main Methods:

• Developed MulNet, a scalable multilayer network framework integrating various molecular interaction data.
• Applied MulNet to analyze diverse cancer datasets, including RNA-sequencing (RNA-seq) data.
• Utilized MulNet for both bulk and single-cell RNA-seq data analysis to reconstruct intra- and intercellular communication.

Main Results:

• MulNet accurately identifies gene modules and key regulators, outperforming state-of-the-art methods on cancer datasets.
• Analysis of colon cancer data identified known regulators and a potential therapeutic target, miR-8485.
• Single-cell analysis of head and neck cancer revealed complex fibroblast-malignant cell communication networks.

Conclusions:

• MulNet provides a high-resolution method for reconstructing complex gene regulatory and communication networks.
• The framework enhances understanding of gene expression and cellular interactions in cancer.
• MulNet offers a powerful tool for discovering therapeutic targets and pathways from multi-omics data.