A novel hypergraph model for identifying and prioritizing personalized drivers in cancer

Area of Science:

• Genomics
• Computational Biology
• Oncology

Background:

• Cancer development involves driver mutations conferring growth advantage, while most mutations are passengers.
• Identifying driver genes is key for effective cancer treatments, but patient heterogeneity poses challenges for current methods.
• Existing computational methods often provide cohort-level driver gene lists, neglecting individual patient variations.

Purpose of the Study:

• To introduce a novel computational method, PDRWH, for prioritizing personalized driver genes in cancer.
• To address the challenge of identifying individual-specific driver mutations by considering downstream gene expression impacts.
• To improve tumor stratification and advance personalized cancer treatment strategies.

Main Methods:

• PDRWH prioritizes mutated genes within a single patient.
• It analyzes the impact of these genes on downstream gene expression across patient groups with shared co-mutations and expression profiles.
• The method was evaluated on 16 TCGA cancer datasets and compared against existing cohort-level and individual-level methods.

Main Results:

• PDRWH successfully identified known general and tumor-specific driver genes.
• The method outperformed existing individual-level and cohort-level approaches across five cancer types.
• PDRWH identified both common and rare driver genes, demonstrating its ability to capture diverse mutational patterns.
• Experimental validation confirmed a predicted driver gene's role in promoting tumor cell proliferation.

Conclusions:

• PDRWH offers a robust approach for personalized driver gene identification in cancer.
• Personalized driver profiles generated by PDRWH can enhance tumor stratification and understanding of tumor heterogeneity.
• The findings support PDRWH's potential to significantly contribute to the development of personalized cancer therapies.