COFACTOR

COFACTOR predicts protein function from structural models by threading target structures through specialized template libraries to infer Gene Ontology terms, Enzyme Commission numbers, and ligand-binding sites.

Key Features:

• Input Structural Models: Accepts structural models derived from experimental determination or computational modeling.
• Template-Based Threading: Threads target structures through three specialized template libraries containing proteins annotated with protein-ligand binding interactions, Enzyme Commission numbers, and Gene Ontology terms.
• Function Prediction Categories: Predicts Gene Ontology terms, Enzyme Commission numbers, and ligand-binding sites from structural similarity.
• Scoring Function: Evaluates confidence by integrating global and local structural similarities between the target and template proteins.
• Recent Improvements: Incorporates sequence profile alignments and protein-protein interaction networks to improve annotation for targets lacking close homologs.
• Benchmarking and Performance: Validated by large-scale benchmarking and recognized as the top method for protein-ligand binding site prediction in CASP9.

Scientific Applications:

• Functional Annotation: Assigns Gene Ontology terms and Enzyme Commission numbers to proteins with unknown or poorly characterized functions.
• Ligand-Binding Site Identification: Identifies potential protein-ligand interaction sites for biochemical characterization and drug discovery.
• Function Hypothesis Generation: Provides structural evidence to generate hypotheses about molecular functions and biological processes.

Methodology:

Threading target structures through three specialized template libraries annotated with protein-ligand binding interactions, Enzyme Commission numbers, and Gene Ontology terms; use of sequence profile alignments; integration of protein-protein interaction networks; and a scoring function that integrates global and local structural similarities.