TGBWDriver: A Cancer Driver Gene Identification Method Based on GraphSAGE and Bidirectional Weighted Feature...

Area of Science:

• Computational biology
• Genomics
• Cancer research

Background:

• Identifying cancer driver genes is crucial for understanding tumor initiation and developing targeted therapies.
• Existing methods often overlook sample-specific gene functions, limiting their accuracy.
• There is a need for computational frameworks that capture context-dependent gene importance.

Purpose of the Study:

• To propose TGBWDriver, a novel computational method for identifying cancer driver genes.
• To address the limitations of existing methods by incorporating sample-specific gene functional differences.
• To improve the accuracy and stability of cancer driver gene ranking.

Main Methods:

• Integration of a two-layer GraphSAGE model with bidirectional weighted feature aggregation.
• Utilizing an exponential pairwise voting strategy for prioritizing candidate driver genes.
• Systematic experiments on BRCA, LUAD, and PRAD datasets to evaluate performance.

Main Results:

• TGBWDriver significantly outperformed five existing methods in precision, recall, and F1-score.
• Ablation studies validated the essential contribution of each component within TGBWDriver.
• The method demonstrated effectiveness in identifying novel cancer driver genes with significant biological relevance.

Conclusions:

• TGBWDriver offers an effective computational framework for accurate cancer driver gene identification.
• The method's ability to capture context-dependent gene functions enhances biological insights.
• This approach holds promise for advancing cancer research and therapeutic strategies.