( Oct 15, 2008: Template set and predictions have been updated. New prediction filtering steps added. )
Previous version: gordion is redirected


MENU

Predictions

Here you can find information related to the predicted protein interactions. The algorithm used for this purpose principally seeks for pairs of polypeptide chains that may potentially interact in a target dataset of protein structures by comparing them with a template dataset of protein interfaces, which is a structural and evolutionary representative of all biological and crystal interactions present in the PDB. If, after comparisons, two structures (be monomeric or complex) are found to structurally and evolutionarily complement each other as chains of any representative interface do (i.e. they are found to match opposite partner chains of an interface), they qualify as a potentially interacting pair. [Learn More...]



Browse matchings grouped by template interfaces [View]

You can browse through our precomputed similarity matchings. A particular matching is between a template interface and the extracted surface of a target structure. The results are grouped by template interfaces. For each one of the 158 template interface a list of structures mathcing to the left and right sides of the template interface is given.

Find a specific match
You can retrieve a set of matchings for a particular template interface and/or a target structure by using the search form below.

Search for a specific binding site match between a template interface and a target structure (input PDB codes).
Template Interface vs. Target Structure
Filter fitty templates.
 

You can also view a list of all the precomputed matchings via this link in a tabular form.


Search for probable interactions

See a list of predictions of protein-protein interactions in which a particular protein in our target dataset is found to participate.

If you know the PDB ID of the protein just type it below.
PDB ID :
Filter fitty templates.

If you have the sequence of the protein first do alignment using BLAST.
Choose a database to search

Enter sequence below in FASTA format


Or load it from disk        

E-value



Online Prediction Calculation

Once all the possible predictions within the target dataset are found, it becomes feasible to predict the binding partners of an input protein. The online computation builds on top of our previous computations.[View]

Below you can try our prediction algorithm by entering the four letter PDB code of a protein or uploading your own structure file that is in the PDB format.

Run our prediction algorithm for a particular input protein that is not in our original target dataset.
Enter the four letter PDB code for automatic download from PDB:

Or load your structure file from disk :       


Walkthrough of this page
help



Mini Help
When the cursor changes to click to get context help.
P.R.I.S.M. : PRotein Interactions by Structural Matching (v 2.00)