PIBASE is a collection of all protein structural interfaces extracted from the Protein Data Bank and PQS structure databases. Both chain-chain and domain-domain (SCOP and CATH definitions) interfaces are detected. A series of properties are then computed for each domain, interface, and complex. The database schema is available here. PIBASE is developed at http://fredpdavis.com/pibase/, with a mirror maintained at https://salilab.org/pibase.

This web interface provides basic search functionality for the database - allowing complex-, interface-, and domain- centric searches. More sophisticated searches can be done by installing a local version of the database and querying it via mysql or the provided perl interface.

Rasmol scripts are provided for viewing the individual interfaces. This script can be used as a rasmol script viewer (in UNIX). The rasmol scripts may also be run on windows machines using "raswin.exe -script scriptname". For help on setting up rasmol as a browser helper please follow this external link.

The interfaces stored in PIBASE have been used to predict protein-protein interactions both within species (S cerevisiae, available at MODBASE; Davis, et al. Nucleic Acids Res 2006) and between species (human and 10 pathogens, available here; Davis, et al. Protein Sci. 2007).

• LGL (Adai, et al. J Mol Biol 2004) is used for all graph layouts. The LGL coordinates are written out to postscript and then converted to PNG using ImageMagick.
• MODELLER ( Sali and Blundell. J Mol Biol 1994) is used to compute solvent accessible surface areas
• DSSP (Kabsch and Sander. Biopolymers 1983) is used for secondary structure assignment.

PIBASE: a comprehensive database of structurally defined protein interfaces. Davis FP, Sali A. Bioinformatics (2005) 21(0): 1901-1907.link. Supplementary Information.