diffusion / fex peptide problem

From: Benjamin Stauch (stauch_at_embl.de)
Date: Wed May 06 2009 - 10:24:16 CDT

Dear community,

I plan to run quite long (hundreds of nanoseconds) simulations of a
small peptide. In the first attempts, the peptide quickly leaves the
center of the solvent sphere (I use spherical boundary conditions) and
diffuses to the surface, resulting in partial "vacuum" contact of the
As I don't want this to happen, and the sphere already has plenty of
cushion, and I expect molecules of that size to diffuse quickly, I would
like to ask if there might be a simple workaround to that problem. Funny
enough, I was not really able to find something useful in documentation
and mailing list archive (either it's too obvious or does not happen to
proteins that easily).
I know that removing the velocity of the center of mass of the sphere
only works on its initial velocity and thus will not help. Is it worth
trying to fix a single atom in the peptide (still calculating forces on
it), as the sphere center itself is defined and should not move? Or
would applying periodic boundary conditions circumvent the problem? I'm
a bit hesitant to do that as I'm not sure about the wrapping of the
individual atoms; I need to work on the molecule's cartesian
coordinates, so in case of periodic boundary conditions I would better
not wrap them in order to not disrupt the molecule?

Thanks and all the best,

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:52:45 CST