TMhhcp

TMhhcp predicts residue-residue contacts and interacting helix pairs in alpha-helix transmembrane proteins to support structural characterization of membrane proteins.

Key Features:

• Residue-residue contact prediction: Predicts residue-residue contacts within alpha-helix transmembrane proteins.
• Interacting helical pair identification: Identifies interacting transmembrane helix pairs.
• Algorithm: Implements Random Forest (RF) models for prediction.
• Evaluation metrics: Reports top L/5 prediction accuracies of 49.5% and 48.8% for two residue contact definitions.
• Pair-level performance: Reports Matthew's correlation coefficients of 0.430 and 0.424 for the two residue contact definitions when identifying interacting helical pairs.
• Benchmarking: Demonstrates superior prediction performance compared to TMHcon and MEMPACK.
• Validation: Performance established using rigorous cross-validation tests.

Scientific Applications:

• Membrane protein contact mapping: Provides contact maps for alpha-helix transmembrane proteins to inform structural models.
• Helix-helix interaction analysis: Identifies interacting helix pairs to aid analysis of helix packing and interface residues.
• Structural annotation: Supports structural characterization of integral membrane proteins that are underrepresented in the Protein Data Bank.
• Functional inference: Contributes contact-derived insights relevant to membrane protein function and interactions.

Methodology:

Random Forest (RF) models were used and performance was evaluated by rigorous cross-validation using two residue contact definitions with top L/5 accuracies and Matthew's correlation coefficient metrics.