X-CAP

X-CAP predicts the pathogenicity of single-nucleotide stopgain variants to aid interpretation of monogenic human disease–associated mutations.

Key Features:

• Target variant class: Predicts single-nucleotide stopgain variants, the third-largest class of mutations associated with monogenic human diseases.
• Machine learning model: Employs a Gradient Boosting Tree (GBT) model trained with a novel methodology and a unique set of features.
• Performance improvements: Achieves an 18% increase in AUROC and a fourfold reduction in false-positive rate relative to existing predictors on large variant databases.
• Clinical prioritization: Optimized for high-sensitivity prioritization of causal stopgain variants in patient exomes for clinical analyses.
• Input format: Operates on Variant Call Format (VCF) files containing genetic variants.

Scientific Applications:

• Clinical diagnostics: Prioritizes candidate pathogenic stopgain variants for interpretation in monogenic disease diagnostics.
• Genomic research: Enables studies of stopgain variant pathogenicity and variant effect interpretation in patient exomes.
• Benchmarking: Serves as a comparator for evaluating variant pathogenicity predictors using large variant databases.

Methodology:

Uses a Gradient Boosting Tree (GBT) model trained with a novel methodology and a unique set of features.