Pred-MutHTP

Pred-MutHTP predicts the pathogenicity of missense variants in transmembrane proteins to discriminate disease-causing from neutral mutations for variant interpretation.

Key Features:

• Sequence-based classification: Implements sequence-based computational methods to classify missense variants in transmembrane protein sequences.
• Training dataset: Trained on a comprehensive dataset comprising 11,846 known disease-causing mutations and 9,533 neutral mutations.
• Evolutionary information: Utilizes evolutionary conservation data to assess the impact of mutations.
• Physicochemical properties: Considers amino-acid physicochemical characteristics and their changes due to mutations.
• Neighboring residue information: Analyzes the influence of adjacent residues on mutation effects.
• Specialized membrane protein attributes: Incorporates number of transmembrane segments, substitution matrices tailored for membrane proteins, and residue distributions in distinct topological regions.

Scientific Applications:

• Variant pathogenicity prediction: Classification of missense variants in transmembrane proteins to distinguish disease-causing from neutral changes.
• Variant prioritization and interpretation: Prioritization of missense variants for experimental validation and interpretation of sequencing data in membrane protein genes.

Methodology:

Integrates evolutionary conservation, amino-acid physicochemical property changes, neighboring residue context, number of transmembrane segments, membrane-specific substitution matrices, and residue distributions across topological regions; models trained on 11,846 disease-causing and 9,533 neutral mutations.