From: Axel Kohlmeyer (
Date: Wed Jul 17 2013 - 02:26:41 CDT

On Tue, Jul 16, 2013 at 5:53 PM, niladri patra <> wrote:

> Hi,
> I need a script to do the following job:
> I have a system with 5000 atoms and it forms several small clusters
> (number of carbon atoms 100 - 200 ) after simulations. Now I want to
> analyze the distribution of cluster sizes (based on carbon numbers) in the
> system.
> I was thinking if I can save a particular cluster type (based on atom
> numbers) from the last frame in a separate pdb file and then I can count
> the number of clusters visually.
> Is there any script available for this job?

cluster analysis can be a tricky issue. finding an algorithmic expression
for which entities are related and which not is an active field of research
in mathematics and computer science and related fields. specifically with
the social media craze and the resulting data available to be mined, there
is a lot of activity recently.

that being said, it all depends on how easily you can set a condition for
which atoms belong to a cluster and which not. the simplest possible
criterion would be distance (or rather distance in relation to radius),
which is what VMD uses - for example - to guess bonds and determine the
fragment id at startup for file formats that don't contain bond

if that would be a sufficient strategy, then you could just set the radius
of your carbon atoms to a suitable value and tell VMD to recompute bonds.
that will also re-evaluate the fragment id and with that you have the
number of fragements and the number for atoms per fragment through some
simple list manipulations in Tcl script.


> Thanks,
> Niladri Patra
> --
> ----------------------------------
> PhD candidate
> Department of Chemistry
> University of Illinois at Chicago
> 845 W Taylor St., Room 4500
> Chicago, IL 60607
> Tel. (312) 996-5393
> ------------------------------------------

Dr. Axel Kohlmeyer
International Centre for Theoretical Physics, Trieste. Italy.