inpPDH

inpPDH predicts DNA-binding hot spot residues in protein–DNA interfaces to identify residues that contribute disproportionately to binding free energy and to inform interpretation of protein–DNA interactions.

Key Features:

• Prediction target: Predicts DNA-binding hot spot residues within protein–DNA interfaces that make disproportionate contributions to binding free energy.
• Hybrid features: Uses hybrid features that combine traditional descriptors with novel interfacial neighbor properties.
• Interfacial neighbor properties: Incorporates interfacial neighbor properties identified as effective in distinguishing hot spot residues.
• Feature selection: Applies a two-step feature selection strategy to remove redundant and irrelevant features.
• Machine learning model: Builds a support vector machine (SVM) predictive model using a subset of seven optimal features.
• Validation: Validated on benchmark datasets with improved prediction accuracy relative to previously published methods.

Scientific Applications:

• Understanding interactions: Supports analysis of molecular mechanisms of protein–DNA recognition and binding.
• Drug design: Informs drug design efforts targeting protein–DNA interfaces by identifying key energetic residues.
• Research insights: Provides insights into protein–DNA interactions relevant to basic research and applied biomedical sciences.

Methodology:

Computational steps include extraction of hybrid features (traditional descriptors plus interfacial neighbor properties), a two-step feature selection to identify seven optimal features, and construction of a support vector machine (SVM) model validated on benchmark datasets.