Inferring Differential Networks by Integrating Gene Expression Data With Additional Knowledge.

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Gene network rewiring is implicated in disease and cell differentiation.
• High-throughput gene expression data enables differential network inference.
• Existing methods struggle with multi-platform data diversity and scRNA-seq limitations.

Purpose of the Study:

• To develop a robust computational framework for inferring differential gene networks.
• To integrate multi-platform gene expression data and prior biological knowledge.
• To address challenges posed by data sparsity, noise, and missing values.

Main Methods:

• Proposed the weighted joint sparse penalized D-trace model (WJSDM).
• Employed a non-paranormal graphical model for data with missing values.
• Utilized a weighted group bridge penalty for integrating multi-platform data and biological knowledge.

Main Results:

• WJSDM effectively infers differential networks from synthetic data.
• Applied to ovarian cancer data, revealing insights into platinum resistance.
• Applied to prostate cancer scRNA-seq data, uncovering mechanisms of anti-androgen resistance.

Conclusions:

• WJSDM provides a powerful approach for differential gene network inference.
• The framework successfully integrates diverse data types and prior knowledge.
• The study yields significant biological insights into cancer drug resistance.