TriOmicNet: A Novel Multi-layer Network Diffusion Approach Integrating Three Types of Scores to Identify Cancer...

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Cancer is driven by genetic mutations, necessitating identification of driver genes for biomarker discovery and personalized therapies.
• Current methods often rely on single data sources, limiting comprehensive driver gene identification.

Purpose of the Study:

• To develop and evaluate TriOmicNet, an integrative multi-layer network diffusion method for prioritizing cancer driver genes.
• To compare TriOmicNet's performance against state-of-the-art methods using multiple data types.

Main Methods:

• TriOmicNet integrates gene mutation, gene expression, and miRNA expression data using a multi-layer network diffusion approach.
• It employs three scoring mechanisms: regulatory potential, control ability, and multi-network diffusion, with a node-scoring strategy for random walk seed nodes.

Main Results:

• TriOmicNet demonstrates competitive performance across various metrics compared to seven state-of-the-art methods.
• The method identified numerous potential driver genes not present in benchmark databases across BRCA, LUAD, and PRAD datasets.
• Ablation studies revealed the dominant contribution of multi-network diffusion score, with other scores offering complementary insights.

Conclusions:

• TriOmicNet offers an effective integrative framework for cancer driver gene prioritization.
• The identified novel driver genes warrant further investigation as potential cancer biomarkers and therapeutic targets.