cardioboost-cardiomyopathies-hg38

cardioboost-cardiomyopathies-hg38 0.0.1

CardioBoost is a disease-specific variant pathogenicity prediction tool designed to assess the likelihood that rare missense variants in genes linked to inherited cardiac conditions (ICCs) are pathogenic. Hosted at https://www.cardiodb.org/cardioboost/, this machine learning classifier, developed by researchers at Imperial College London and partners, focuses on cardiomyopathies (e.g., hypertrophic and dilated cardiomyopathy) and inherited arrhythmia syndromes (e.g., long QT syndrome and Brugada syndrome). CardioBoost outperforms general genome-wide tools by providing higher accuracy—91% for cardiomyopathies and 96% for arrhythmias (precision-recall AUC)—particularly for variants classified with over 90% confidence. Users can input a gene and variant (in cDNA format) to obtain pathogenicity probabilities, aiding variant annotation for research or clinical interpretation. Precomputed scores for all possible rare missense variants in ICC-related genes are available, though it’s not a standalone clinical decision tool and should complement ACMG guidelines. The tool’s data and source code are accessible on GitHub, supporting transparency and further analysis.