From: Mark Cunningham (
Date: Wed May 15 2013 - 10:58:17 CDT


Attached is a tcl script that will accomplish what you want. It will work with large dcd files
but it takes a long time to run (single-threaded and interpreted).
You're best off running it remotely as -dispdev text.

The function draw_corr will also read the output matrix file and connect CA carbons (above
a threshold) that display correlated motion. I found this to be useful.

I used Matlab to plot the resulting correlation images, and have attached some .m files that will
make matrix plots. (The main file is ccplot.m.) Results are illustrated in a paper I wrote
(J. Mol. Modeling 18 (2012) 3147-3152)

You'll have to mold this to your specific use but it should give you a starting point.


From: [] on behalf of Prof. Eddie []
Sent: Tuesday, May 14, 2013 11:12 AM
Subject: vmd-l: dynamic cross correlation map

Hi all,
Has anyone written a Dynamic cross correlation algorithm for vmd? It does not seem to have one and I can't find any on the web. I thought I'd ask before writing my own.

I'd like a measure of how correlated the motion of one residue is compared with every other residue.
I think what I'm looking specifically is a function that could (for each frame or average over all time frames) calculate C_ij = <r_i*r_j>/sqrt[<r_i^2><r_i^2> ].


Edward Ackad, Ph.D<>
Assistant Professor of Physics
Computational Nanophotonics
Southern Illinois University Edwardsville
(618) 650-2390

  • application/mathematica attachment: ccplot.m
  • application/mathematica attachment: rpimage.m