[Chimera-users] clustering of docking solutions

Elaine Meng meng at cgl.ucsf.edu
Mon Nov 17 10:18:39 PST 2008

Hi Sebastian,
Are all the docked structures the same molecule, just in different  

You are right that ViewDock doesn't have clustering, and Ensemble  
Cluster automatically determines the clusters without allowing you to  
specify an RMSD cutoff.

My ideas only pertain if the structures are all the same molecule or  
at least contain a common set of atoms:

(a) You could use Ensemble Match to compare the ensemble to itself, if  
the atoms in common have the same names.  This would not perform  
clustering, but it calculates all-by-all RMSD values using a set of  
atoms you specify.  However, these are best-fit RMSDs of the  
structures, while I suspect you may want the RMSDs in their current  
positions instead.  Another caveat is that it creates a dialog that is  
an N x N table, which may not be manageable for a large number of  
structures (high N).
There is an example of using Ensemble Match at the end of the "model  
panel and ensembles" tutorial:

(b) As you mentioned, you could convert to a multi-model PDB, and use  
the MD Movie tool's RMSD analysis to calculate an all-by-all RMSD  
map.  However, it would require more work to extract the RMSD values  
(output is a grayscale map) and these are also best-fit RMSDs.  This  
option is more practical for high N than option (a), however.
Part 2 of the "trajectories/ensembles" tutorial includes an example of  
MD Movie RMSD Analysis:

(c) You could use a script to do all-by-all comparisons of the  
structures with the "rmsd" command.  This calculates current-position  
RMSDs rather than best-fit RMSDs.

See these previous posts for more general information on scripts:

I hope this helps,
Elaine C. Meng, Ph.D.                          meng at cgl.ucsf.edu
UCSF Computer Graphics Lab (Chimera team) and Babbitt Lab
Department of Pharmaceutical Chemistry
University of California, San Francisco

On Nov 17, 2008, at 6:07 AM, Sebastian Kruggel wrote:

> Hi all,
> I am looking for a possibility to cluster docking results (multimol2
> file) in chimera. In DockView I can't find such tool and in the
> EnsembleCluster method I don't see a way to set rmsd limits.
> I found a hint to the MD Analysis tool in the mailing list, ok, I  
> could
> convert my docking solutions into pdbs and use this, but probably  
> there
> is another simple and convenient way that I am just too blind to  
> see ;-)
> Maybe somebody can help?
> Thanks in advance,
> Sebastian

More information about the Chimera-users mailing list