driveR

driveR prioritizes cancer driver genes from somatic genomics data using a multi-task learning framework that supports both personalized (patient-level) and cohort-scale analyses.

Key Features:

• Personalized and Batch Analysis: Performs individualized (patient-level) and cohort-level (batch) prioritization of cancer driver genes from somatic genomics data.
• Integration of Genomic Information and Biological Knowledge: Combines comprehensive genomic features with prior biological knowledge to inform and refine driver gene prioritization.
• Multi-Task Learning Model: Implements a multi-task learning model to capture shared patterns across different cancer types and patient samples.
• Performance and Validation: Validated on 28 diverse datasets and shown to outperform existing methods in prioritizing driver genes.

Scientific Applications:

• Cancer Research: Prioritizes driver genes to identify genetic alterations underlying tumorigenesis and to inform targeted therapy research.
• Genomic Data Analysis: Applies to both individual patient somatic genomic profiles and larger cohort datasets, including studies of rare cancers or specific subtypes.

Methodology:

Combines personalized and batch analysis, integrates genomic information with biological knowledge, and applies a multi-task learning model, with validation performed on 28 datasets.