AI-Driver

AI-Driver predicts the driver status of somatic missense mutations in cancer genomes by using an ensemble classifier that integrates 23 pathogenicity features to distinguish drivers from passengers.

Key Features:

• Ensemble Methodology: Integrates multiple predictive models to assess driver status of somatic missense mutations.
• Pathogenicity Feature Integration: Utilizes 23 distinct pathogenicity features to capture multifaceted mutation impacts.
• Cancer-Specific Predictions: Produces mutation-level predictions tailored to personal cancer genomes.
• Benchmark Performance: Demonstrated superior and stable performance across four independent benchmarks.
• Pre-computed Variant Scores: Includes pre-computed AI-Driver scores for all possible human missense variants.

Scientific Applications:

• Driver Mutation Identification: Prioritizes somatic missense mutations as drivers or passengers in personal cancer genomes and exome sequencing data.
• Therapeutic Target Discovery: Supports identification of candidate driver alterations that may inform targeted therapy development.
• Tumor Biomarker Prediction: Contributes to prediction of tumor biomarkers applicable to diagnostics and monitoring, including liquid biopsy approaches.

Methodology:

AI-Driver uses an ensemble learning approach that combines multiple predictive models and integrates 23 pathogenicity features to classify somatic missense mutations as drivers or passengers.