Rhapsody

Rhapsody predicts the pathogenicity of missense variants by integrating protein dynamics, structural properties, sequence conservation, and Pfam-derived coevolutionary information within a machine learning framework.

Key Features:

• Protein Dynamics Integration: Incorporates dynamic structural properties of proteins to improve prediction of variant pathogenicity.
• Pfam Coevolutionary Data: Utilizes coevolutionary information derived from Pfam multiple sequence alignments to contextualize amino acid substitutions.
• In Silico Saturation Mutagenesis: Performs systematic evaluation of all 19 possible amino acid substitutions at each residue position in human proteins.
• Large-Scale Variant Benchmarking: Evaluates predictive performance using datasets containing approximately 20,000 annotated variants.

Scientific Applications:

• Variant Pathogenicity Prediction: Assesses the functional impact of missense mutations associated with human diseases.
• Clinical Genomic Interpretation: Supports analysis of disease-associated variants in clinical molecular diagnostics.
• Functional Variant Analysis: Investigates mutation effects through saturation mutagenesis across protein sequences.

Methodology:

Rhapsody applies a machine learning framework that integrates sequence conservation, protein structural properties, protein dynamics, and Pfam-derived coevolutionary data to predict the pathogenicity of missense variants.