The Research Collaboratory for Structural Bioinformatics (RCSB) has opened both the new Protein Data Bank (PDB) Web site and ftp archive. On February 3rd 1999, the RCSB organization accordingly has assumed the primary responsibility for distributing new entries to the PDB. Following the first five successful updates, the PDB now contains 9419 coordinate entries for proteins, nucleic acids, and carbohydrates.
During the first two weeks of operation in February, the RCSB PDB Web site recorded an average of 3682 hits per day. In turn, the ftp site which contains the structural data, each day transmitted an average of 5040 files. Since RCSB officially became the primary distributor of new structural data, the RCSB PDB has received approximately 50 new entries each week. The new entries are processed by software developed by the RCSB while the throughput has kept good pace with the deposition rate for new structures.
The opening of the new Web and ftp sites marks a very important step in the management transition of the Protein Data Bank from Brookhaven National Laboratory to the RCSB, a non-profit organization consisting of groups from Rutgers University, San Diego Supercomputer Center (SDSC), and NIST, the National Institute of Standards and Technology. PDB already exists 26 years and constitutes the single world repository for atomic structures of biological macromolecules.
A highlight of the new site is SearchLite, a search engine developed at SDSC, which represents the first phase in the development of new tools for querying the data bank. This tool allows the user to carry out keyword-based searches across the annotated entries in the entire database with several options for the analysis, display, and download of the coordinates for the molecules of interest.
Another key service provided by the RCSB is a validation server developed at Rutgers. This server can be used by structure depositors in order to detect possible errors in the atomic models before final submission to the PDB. Not only does this server provide the ability to check the stereo-chemical quality of the models, but more critically, it also implements methods for checking the agreement of the atomic model with crystallographic data.
Other Web services maintained by the RCSB PDB include extensive lists of on-line resources, a discussion group for topics related to protein structure, and the PDB newsletter, available on-line or by e-mail subscription. For more information, please consult the RCSB PDB Web site, as well as the VMW article Five year funding will significantly extend Protein Data Bank's capabilities which appeared in the December 1998 issue.