NetWalkRank: Cancer Driver Gene Prioritization in Multiplex Gene Regulatory Networks by a Random Walk Approach

Area of Science:

• Oncology
• Bioinformatics
• Systems Biology

Background:

• Identifying cancer driver genes (CDGs) is essential for understanding cancer development.
• Existing methods often struggle to integrate complex, multi-stage gene information.
• Multiplex gene regulatory networks (GRNs) offer a promising approach to model gene interactions.

Purpose of the Study:

• To develop a novel network-based framework, NetWalkRank, for prioritizing CDGs.
• To leverage multiplex GRNs and gene expression data for enhanced CDG identification.
• To validate NetWalkRank's effectiveness in prioritizing hepatocellular carcinoma (HCC) driver genes.

Main Methods:

• Constructing multiplex gene regulatory networks (GRNs) integrating multi-stage gene information.
• Applying network propagation within the multiplex GRNs to assess gene abnormality spread.
• Utilizing gene expression profiling data as input for the network analysis.
• Training a random forest model with NetWalkRank scores for CDG prediction.

Main Results:

• NetWalkRank effectively prioritized known CDGs for hepatocellular carcinoma (HCC).
• The framework demonstrated superior performance compared to existing driver gene ranking methods.
• A random forest model trained on NetWalkRank scores achieved accurate CDG prediction.
• Numerical experiments confirmed the efficiency and effectiveness of the proposed method.

Conclusions:

• NetWalkRank provides a robust framework for prioritizing cancer driver genes.
• Integrating information across multiplex GRNs significantly enhances CDG identification and prediction.
• The method holds promise for advancing cancer research and therapeutic strategies.