PIBASE: a comprehensive database of structurally defined protein interfaces

• DOI: 10.1093/bioinformatics/bti277
• Publication Date: 2005-01-19T02:58:09Z
• Authors: F. P. Davis; A. Sali
• Keywords: Interface (matter); Protein Data Bank; Protein structure database

Motivation: In recent years, the Protein Data Bank (PDB) has experienced rapid growth. To maximize utility of high resolution protein–protein interaction data stored in PDB, we have developed PIBASE, a comprehensive relational database structurally defined interfaces between pairs protein domains. It is composed binary extracted from structures PDB and Probable Quaternary Structure server using domain assignments Structural Classification Proteins CATH fold classification systems. Results: PIBASE currently contains 158 915 interacting 105 061 domains 2125 SCOP families. A diverse set geometric, physiochemical topologic properties are calculated for each complex, its domains, binding sites. subset interface used to remove redundancy within entries, resulting 20 912 distinct domain–domain interfaces. The complexes grouped into 989 topological classes based on their patterns contacts. corresponding sites categorized 18 755 30 975 classes, respectively, topology secondary structure elements. illustrated by outlining several current applications. Availability: accessible via world wide web at http://salilab.org/pibase Contact: [e-mail] Supplementary information: http://salilab.org/pibase/suppinfo.html