Zenpatches is a database containing protein-protein contact surface
predictions of the entire PDB ensemble, generated using BioDec's Zendock
For every PDB file there is a prediction of the whole protein and of its
separated chains (the prediction may use the surface information derived
from the existing PDB chain contacts).
Predicted protein contacts are grouped into subsets that could be seen as
"Patches", from the point of view of the protein surface. Each Patch is
obtained grouping the predicted contacts according to a C-alpha distance
threshold of 4.5 Angstrom. This measure is calculated averaging the minimum
C-alpha distances between surface amino acids of a very large representative
protein ensemble and then adding to it twice the value of the overall
The Zenpatches output is represented by a result page showing information on
the protein as they can be extracted from the pdb file, showing information
on the predicted patches, giving an image of the protein and its highlighted
patch parts and finally giving links to some other useful data (i.e. links
to the pdb file, to the rcsb pdb page, ....).
- Protein Data
- Protein information: with protein name and chains.
- Title: as extracted from the PDB "TITLE" field.
- Classification: as derived from the PDB "HEADER" field. It has also
a link to the "htdig" search engine: a new search with the
"Classification" keywords is run.
- Compound: data extracted from the PDB "COMPND" fields. It contains
type classification of the molecules, their names and finally their
relative chain assignation.
- Exp. Method: the source of the experimental data of the protein
structure, extraceted from the PDB "EXPDTA" field.
- Source: the Scientific Organism source of each PDB molecule type and
eventually other data on fragmentation and on the expression system
(as extracted from the PDB "SOURCE" fields).
- Protein Links
- PDB (file): it's the link to the PDB file.
- PDB (@rcsb.org): it's a link to the Protein data Bank file page
at rcsb (www.rcsb.org).
- VRML: it's a link to the VRML file of the prediction. You must have
the right application to open it, (a decompression program, as "gzip"
or "winzip" and a VRML viewer). Only VRML 2.0 is generated.
- MolMol script: it's a link to the Molmol script of the prediction.
There are not many Molmol viewers plugins for web browsers, so you
must have Molmol installed on your machine to see it working.
- Protein Prediction
- Patches Table: It's a table showing each group of surface contacts
(patch), gathered by a C-alpha distance threshold. Each patch has the
list of the PDB residues predicted to be in contact, specifying their
PDB number and their name. Enclosed within brackets there is the
reliability of the predicted contact, that spans a range from 0 to
10. The reliability is a function of the ZenDock predictor
output, and a higher value is a higher probability that the
predicted contact is correct.
The probability that a predicted contact is really a contact grows
almost linearly from 70% to 90% with respect to an increase in the
reliability value. In bold font there are prediction with a
reliability equal to or higher to a threshold of 3.
- Patches Image: It's an image representing protein in the "cartoon"
format. Patches are highlighted as yellow surfaces. Only contacts
with a reliability equal to or higher than a threshold of 3 are
shown (i.e. the ones displayed in bold font in the Patches Table).