EPAD

EPAD models position-specific, distance-dependent statistical potentials for proteins by parameterizing observed atomic interaction probabilities using atom types, sequence profile context, radius of gyration, and evolutionary information to improve protein-structure assessment.

Key Features:

• Position-Specific Parameterization: Considers the protein sequence profile context for each atom pair, producing distinct energy profiles for identical atom pairs depending on their sequence position.
• Distance-Dependent Modeling: Employs a distance-dependent framework that accounts for spatial relationships between atoms in computing potentials.
• Atom-Type-Based Interaction Probabilities: Parameterizes observed atomic interacting probabilities by atom types rather than relying solely on simple counting methods.
• Radius of Gyration Incorporation: Integrates the radius of gyration as a structural descriptor in the parameterization of interaction probabilities.
• Evolutionary Information Integration: Incorporates evolutionary data to enhance the reference state and improve the discriminative performance of the potentials.
• Knowledge-Based Potentials: Derives energy potentials using the inverse of the Boltzmann law as a knowledge-based approach.

Scientific Applications:

• Decoy Discrimination: Demonstrates superior performance in discriminating decoy from native-like protein structures in benchmark tests.
• Protein Structure Prediction: Provides position-specific energy profiles useful for evaluating and ranking predicted protein models.
• Functional Annotation: Aids functional annotation by refining energetic assessments that relate to structural plausibility.
• Protein Folding Studies: Supports studies of folding mechanisms by helping to distinguish correct versus incorrect structural conformations.

Methodology:

EPAD parameterizes observed atomic interacting probabilities by combining atom types with sequence profile context and radius of gyration within a distance-dependent framework, derives knowledge-based potentials from the inverse Boltzmann law, and enhances the reference state using evolutionary information.