Identification of cancer driver genes based on hierarchical weak consensus model

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Cancer arises from accumulated gene mutations during somatic cell evolution.
• High-throughput omics technologies generate vast datasets, posing challenges for identifying cancer driver genes.
• Existing methods like frequency-based and network-based approaches have limitations in identifying low-mutation-rate drivers and integrating feature relationships.

Purpose of the Study:

• To propose a novel method for identifying cancer driver genes by effectively integrating multi-omics data and feature interconnections.
• To introduce a new topological feature, the co-mutation clustering coefficient (CMCC), derived from PPI and co-mutation hypergraph networks.
• To develop the Hierarchical Weak Consensus (HWC) model for driver gene identification.

Main Methods:

• Analysis of existing multi-omics data integration methods.
• Development of the co-mutation clustering coefficient (CMCC) by analyzing PPI and co-mutation hypergraph networks.
• Integration of CMCC, mRNA, and miRNA differential expression scores using a hierarchical weak consensus model to create the HWC method.

Main Results:

• The HWC method demonstrated superior performance compared to seven state-of-the-art methods across three cancer types.
• HWC achieved better statistical evaluation index, functional consistency, and partial area under the ROC curve.
• The proposed method effectively addresses limitations of previous frequency-based and network-based driver gene identification techniques.

Conclusions:

• The Hierarchical Weak Consensus (HWC) model offers a robust and effective approach for identifying cancer driver genes.
• Integrating novel topological features like CMCC with multi-omics data enhances driver gene discovery.
• HWC represents a significant advancement in computational oncology for understanding cancer gene mutations.