eF-seek

eF-seek predicts protein functional sites by comparing electrostatic potential and molecular surface shape to identify ligand-binding and active sites from three-dimensional structures.

Key Features:

• Database Utilization: Searches the eF-site database, including eF-site/antibody, eF-site/prosite, eF-site/P-site, and eF-site/ActiveSite subsets, for characteristic ligand-binding surfaces.
• Electrostatic and Shape Analysis: Calculates electrostatic potentials by numerically solving the Poisson-Boltzmann equation on molecular surfaces and compares surface shapes while incorporating residue hydrophobicity.
• Graph Theory Application: Employs a clique-detection algorithm from graph theory to search and discriminate characteristic molecular surfaces within the eF-site database.
• Molecular Surface Computation: Computes molecular surfaces from atom coordinates and annotates them with physical properties for comparison.
• Input/Output Formats: Operates on PDB format coordinate files and outputs predicted virtual complex structures in PDB format.

Scientific Applications:

• Structural genomics and proteomics: Infers the geometrical location of functional sites in hypothetical or unannotated proteins using surface and electrostatic similarity.
• Ligand- and phosphate-binding site identification: Identifies ligand-binding sites, including phosphate-binding (P-site) and antigen/antibody interfaces, by matching to known eF-site entries.
• Functional annotation of new structures: Predicts active sites with functions similar to known proteins to assist annotation of newly determined protein structures.

Methodology:

The method expands searches to heteroatom binding sites in the Protein Data Bank and determines similarity thresholds using solved free and complex structures; it computes molecular surfaces from atom coordinates annotated with physical properties, numerically solves the Poisson-Boltzmann equation to derive electrostatic potentials, incorporates residue hydrophobicity, applies clique-detection search against the eF-site database, and generates virtual complex structures in PDB format.