Summary of Antibody Structures in the Protein Databank, automatically updated each month.
Introduction
This page provides an automatically updated summary of all antibody structures in the Protein Databank
Given the problems with parsing out useful information from PDB files in an automated fashion, there may be some errors in these summaries. We would be grateful if users would notify us if any such problems are found!
Methods
Old Method (2002)
We mirror PDB files and use a Make based system for automatic processing of new PDB files as they arrive.
As part of this processing, a Perl script is used to determine whether a file is an antibody. If so, summary information about that antibody is extracted from the file using Perl and C programs and the results are written in XML. At this stage, manual correction of errors detected in the automatic processing can be performed.
New Method (Autumn 2020)
As before we mirror PDB files locally, but SACS now uses a fully sequence-based method to identify antibodies. This uses a forward search (using known antibody sequences) to identify potential antibodies followed by a reverse search of those potential sequences against a database of known antibodies and T-Cell Receptors to ensure that the sequence is closer to an antibody that to a TCR. This is done using findpdbabs
Both methods
Individual XML files are concatenated into a master file containing information on all the antibodies. XSLT is then used to create a summary view of all antibodies or a detailed view of a single antibody on the fly as you view the pages.
If you use this site, please site: Allcorn, L. C. and Martin, A. C. R. (2002) SACS - A Self-Maintaining Database of Antibody Crystal Structures, Bioinformatics, 18, 175-181. [PDF]