Created by Donald E. Curtis with help from Todd Dolinsky and Nathan Baker
Use PropKa and PDB2PQR to add hydrogens to your x-ray structure: PDB2PQR
Try our mirror site in the US: http://propka.chem.uiowa.edu
The PROPKA method is developed by the
Jensen
Research Group
Department of Chemistry
University of Copenhagen
The PROPKA method is described in Hui Li, Andrew D. Robertson, and Jan H. Jensen "Very Fast Empirical Prediction and Interpretation of Protein pKa Values" Proteins, 2005, 61, 704-721. Please cite this reference in publications. A reprint can be obtained by contacting Jan Jensen.
NMR structures are likely to yield inaccurate pKa values for some residues. Alternatively, PROPKA can be used to validate protein structures. See N. Powers and Jan H. Jensen "Chemically Accurate Protein Structures:Validation of Protein NMR Structures by Comparison of Measured and Predicted pKa Values", J. Biomolecular NMR 2006, 35, 39-51. A preprint can be obtained by contacting Jan Jensen.
While the web interface can access files directly from the PDB, it is a good idea to take a careful look at the PDB file before using it for pKa prediction.
It is currently up to the user to make sure that the PDB file contains the biologically active unit. For example, the unit cell can include more than one protein or exclude chains related by symmetry.
Occasionally more than one conformation for part of the protein is included in the PDB file (with fractional occupancy). If so, you must create separate PDB files for each conformation.
The orientations of Gln, Asn, and His side chains can be difficult to assign experimentally because the density of N, O, and C atoms are difficult to distinguish at low resolution (~1.5 Å and above). In several cases we have found significantly better agreement with experiment after “flipping” such side chains by 180° (i.e. by interchanging the atom labels in the PDB file).
The current version of PROPKA does not include the effect of explicit water molecules, ions, or ligands on the protein pKa values (i.e. all HETATM's are ignored).
Hui Li, Andrew D. Robertson, and Jan H. Jensen "Very Fast Empirical Prediction and Interpretation of Protein pKa Values" Proteins, 2005, 61, 704-721.
N. Powers and Jan H. Jensen "Chemically Accurate Protein Structures:Validation of Protein NMR Structures by Comparison of Measured and Predicted pKa Values", J. Biomolecular NMR 2006, 35, 39-51.
Melissa A. Porter, Jordan R. Hall, James C. Locke, Jan H. Jensen and Pablo A. Molina "Amide Backbone Interactions are the Prime Determinants of Carboxyl pKa Values at N-termini of alpha-helices" Proteins: Structure, Function, and Bioinformatics, 63, 621 - 635.
MN Davies, CP Toseland, DS Moss and DR Flower "Benchmarking pKa Predictions" BMC Biochemistry 2006, 7, 18.
January 18, 2005. PROPKA 1.00 web interface established
Work on the PROPKA method is funded by The Danish Agency for Science, Technology and Innovation (FNU) and a grant from the Marie Curie Actions Program.