Triangle Match - Pairwise Protein Structure Alignment
Select Model:
Input the Model protein to be aligned:
Type PDB code (e.g. 1hc7) and Chain/s (optional) (e.g. AC),
or press "browse" to upload your own .pdb file.
Select Target/s:
Choose between running against a Single Target PDB File,
or running against our DB of 2158 PDBs representing superfamilies.
Select Processing:
Full Processing option is slower, due to considering more triangles, but may yield larger matches.
Fast Processing option is quicker, thanks to considering less triangles, but might yield smaller matches.
You may click show advanced options to set these parameters in attempt to get better results.
|
Set Advanced Options (optional)
|
Voting:
Voting epsilon:
2 model-target triangles are considered "similar", hence their heads constiute a possible model-target atom pair match,
if their sides lengths differ in less then this epsilon.
Default is 1.3 angstrum.
Min votes to qualify:
Triangle bases with at least this num of votes are processed
Triangle bases with less then this num of votes are discarded
Default is 7 votes.
Verifying:
Verification max distance:
The algorithm verifies each long enough vote list,
by discarding in each iteration the "worst" match.
Verification max distance is the lower bound for the verification stage.
The lower it is, the less lists will pass verification.
Default is 1.0 angstrum.
Extending:
Extension type:
You can choose from the following 2 different Extension type:
Hausdorff Extension (smaller matches,lower RMSD,faster)
Bipartite Matching (larger matches,higher RMSD,slower)
Default and recommended is Hausdorff.
Extension epsilon:
Extending and refining in each iteration is done by adding matches that are below this epsilon distance,
and deleting matches that their distance is greater then thi epsilon.
Default is 3.0 Angstrom.
Num of extension iterations:
Too few extension iterations might yield matches that are large, but are not "refined" or "true" matches.
Default is 10 extension iterations.
Clustering:
Max RMSD to cluster:
Extending stage clusters 2 transformations into one if their RMSD is less than this bound.
Default is 0.5 RMSD.
You may set the Number of best matches you would like to see. Maximum 50.
After you have set the Model ans Target/s input proteins, and, optionally, the advanced parameters,
Press the Run! button.
If Allow some time for the program to run.
The results will appear.
The matches found will show, sorted primarily by score, and secondarily RMSD.
Each result line has one link to display the alighnment's details,
and another link to download the alignment's .pdb file.
Bear in mind that in many cases the numerically favorable match is not the one you're looking for.
You'll have to do a little browsing.
The P-value calculation is based on results of an all-agaist-all matching of superfamilies representatives DB that contains 2158 PDB's. This DB is part of our Download
|
Downloading TriangleMatch
|
You may Download - Linux 64bit
the program with our DB.
For Model against DB runs, or for very large PDB inputs, is recommended to download and run locally.
| If you use this program, please cite: |
1. R. Nussinov, H.J. Wolfson, Efficient Detection of Three - Dimensional Motifs in Biological Macromolecules by Computer Vision Techniques,
Proceedings of the National Academy of Sciences, U.S.A., 88, 10495-10499, (1991).
[ Link ]
2. H.J. Wolfson and I. Rigoutsos, Geometric Hashing: An Overview, IEEE Computational Science and Engineering, 4(4), 10-21, (1997).
[ Link ]