A B-factor DataBank (BDB) entry is a PDB entry with full isotropic B-factors.

What is a B-factor?


Protein engineers, homology modellers, biologists, crystallographers and bioinformaticians frequently analyze B-factors in PDB structures because B-factors are a measure of mobility. Normally full B-factors are stored in the B-factor field in the ATOM records of a PDB file. However, in about 10% of the X-ray PDB files the B-factor field represents different quantities. For example, “residual” B-factors without the isotropic TLS contribution can be present, atomic mean-square displacements instead of B-factors have been deposited, or sometimes the anisotropic overall scale has not been included in the B-factors. In general, this leads to lower B-factors than the B-factors reported for the majority of the PDB files. Furthermore, the location of apparent (normalized) B-factor maxima in a chain might be different and might even shift to a different secondary structure element of a different secondary structure type when full (estimates of) isotropic B-factors instead of residual B-factors are considered. Therefore, the different representations of disorder may influence both large and small scale analyses of B-factors. The BDB aims to present a more consistent B-factor representation. The BDB contains PDB files with full isotropic B-factors in the B-factor field if the original PDB file contains enough information to determine the content of the B-factor field and calculate the full B-factor if necessary.


If you use BDB entries, please cite:

Wouter G. Touw & Gert Vriend, “BDB: Databank of PDB files with consistent B-factors” (2014) Protein Engineering, Design and Selection (PEDS) 27 (11): 457-462. doi: 10.1093/protein/gzu044

Download BDB entries

You can go to the download page of a single BDB entry by entering a PDB code in the search box. The entire databank can be downloaded via rsync:

rsync -avz rsync://rsync.cmbi.umcn.nl/bdb/??/????/????.bdb my-bdb/

A list of all BDB entries is available here.

Source code

The source code for generating BDB files from PDB files is available on GitHub.

Explanation of BDB entry pages

These pages will show whether full B-factors have been calculated from the PDB file or if full B-factors were already present in the PDB file. The information that was used to arrive at this conclusion is shown in the Refinement and PDB Entry Format sections. The refinement section is subdivided in software, TLS-related info, other B-factor info and other refinement remarks.

B-factor type according to PDB remediation: in 2011 several problems were remediated by the PDB. The remediations included fixing TLS group definitions and determining whether the B-factor type in the PDB file is likely to be residual or not. If the B-factor type could not be verified according to the remediation a BDB file has not been created.

Number of TLS groups and B-factor type according to refinement remarks: these table rows show the information on the number of TLS groups or the type of B-factor that could be extracted from REMARK 3 records.

SKTTLS validation summary: results of TLS model tests after a TLSANL run. Extreme residuals may indicate ADP problems. A detailed explanation is given in the SKTTLS report description and in this paper about validation of TLS models by Zucker, Champ & Merritt (2010).

B-factor group type: indicates if the protein and/or nucleic acid chain(s) have been refined using a single overall B-factor, one or two B-factors per residue, or individual B-factors.

Cα only: indicates if the protein chain only consists of a Cα trace.

P only: indicates if the nucleic acid chain only consists of a phoshorus trace.

Reproducible fraction of B-factors: the fraction of Beq values in the B-factor field of the ATOM records that could be reproduced using the anisotropic temperature factors in the ANISOU records.

Anisotropic temperature factors specified in correct order: indicates if Beq values could be calculated using the first three elements of the anisotropic temperature factors in the ANISOU records (U11, U22, and U33) or if a non-standard Uij combination was necessary for the calculation.

Other refinement remarks: the “OTHER REFINEMENT REMARKS” section of REMARK 3.

PDB entry format: date and version of the PDB format used for the PDB file.

Created by

Wouter Touw & Gert Vriend