DGRanker: Cancer Driver Gene Detection in Human Transcriptional Regulatory Network

Area of Science:

• Genomics
• Network Science
• Computational Biology

Background:

• Cancer driver genes (CDGs) are crucial for understanding cancer initiation.
• Current CDG identification methods primarily rely on gene expression and genomic mutation data.
• Networking techniques and influence maximization offer new avenues for CDG discovery.

Purpose of the Study:

• To construct a cancer transcriptional regulatory network.
• To identify CDGs using a network science approach, independent of mutation and genomic data.

Main Methods:

• Employ social influence network theory to model gene regulatory networks (GRNs).
• Create GRNs using gene expression data, nodes, and edges.
• Adapt an algorithm by weighting regulatory edges with an influence spread concept.
• Identify CDGs based on node ratings derived from influence and power metrics.

Main Results:

• The proposed network-based method demonstrates superior performance in identifying CDGs compared to existing computational and network approaches.
• The method successfully identifies numerous CDGs that were missed by previously published techniques.
• The approach highlights the efficacy of network science in uncovering critical cancer-related genes.

Conclusions:

• Google's PageRank algorithm can be modified for effective CDG identification within transcriptional regulatory networks.
• This network-based method serves as a valuable complement to existing computational tools for cancer driver gene discovery.