PRODIGY: personalized prioritization of driver genes

Area of Science:

• Oncology
• Computational Biology
• Genomics

Background:

• Cancer evolution is driven by a small subset of somatic mutations impacting cellular processes, leading to tumor development.
• Identifying these critical driver genes is essential for understanding cancer mechanisms and developing targeted therapies.
• Current methods primarily focus on cohort-level analysis, leaving patient-specific driver gene identification a significant challenge.

Purpose of the Study:

• To develop and validate a novel algorithm for patient-specific ranking of cancer driver genes.
• To improve the identification of genes that critically influence tumorigenesis at an individual patient level.

Main Methods:

• Developed PRODIGY, an algorithm that integrates patient expression and mutation profiles with pathway and protein-protein interaction data.
• Employs the prize-collecting Steiner tree model to quantify the impact of mutated genes on deregulated pathways.
• Ranks mutated genes based on their aggregated impact across all identified deregulated pathways.

Main Results:

• PRODIGY demonstrated superior performance in identifying validated driver genes across five TCGA cancer cohorts (>2500 patients) compared to existing methods.
• The algorithm successfully identified rare driver mutations, highlighting the pleiotropic effects of driver genes.
• Results indicate PRODIGY's capability to pinpoint patient-specific drivers, advancing personalized cancer treatment.

Conclusions:

• PRODIGY offers a robust approach for patient-specific driver gene identification, moving closer to personalized cancer medicine.
• The algorithm's ability to detect rare drivers enhances its clinical utility for diverse cancer types.
• PRODIGY facilitates the discovery of actionable targets for individualized cancer therapies.