TMKink

TMKink predicts kinks in transmembrane helices from protein sequence to support analysis of membrane protein structure and function.

Key Features:

• Local Sequence Preferences: Analyzes local sequence context within transmembrane helices and identifies residue preferences such as proline enrichment associated with bends.
• Neural Network-Based Prediction: Applies a neural network classifier to predict kink occurrence, reporting sensitivity of 0.70 and specificity of 0.89.
• High Reliability: Combines sequence-based features and neural network scoring to achieve high specificity and reduce false positives.

Scientific Applications:

• Structural Analysis: Predicts helix kinks to inform membrane protein architecture and structural modeling.
• Functional Insights: Identifies potential kink sites that can influence membrane protein dynamics involved in signaling, transport, and enzymatic activity.
• Protein Engineering: Guides design or modification of membrane proteins by indicating sites prone to helix distortions.

Methodology:

Analyzes local sequence data to assess sequence preferences (notably proline enrichment) and applies a neural network model to predict helix kinks.