VMD-L Mailing List

From: Peter Freddolino (petefred_at_ks.uiuc.edu)
Date: Tue Mar 11 2008 - 07:45:44 CDT

This email seems particularly timely since it's the second that I've
received in the last couple days about the switching issue. When I
started working on namdenergy I'd assumed that people would mostly use
it with PME, as with normal simulations. Since that doesn't seem to be
the case, let me point out the following: The right way to deal with the
cutoff dependence of electrostatic interactions without PME (which is
due to switching) is to turn switching off, *not* to use very large
cutoffs. I've committed a version of namdenergy to cvs that turns
switching off if you feed it a switch distance that is blank (literally
"", not just unspecified) or less than 0. The docs have been changed to
reflect this. For now I've posted this version at
www.ks.uiuc.edu/~petefred/namdenergy.tcl; it will be included in future
vmd releases.

Best,
Peter

poker_at_physics.usyd.edu.au wrote:
> There's been an earlier thread (IIRC,) about interaction energy depending
> on cutoff values used, but..
>
> I'm writing to ask whether people have a version of namdenergy.tcl (or
> another script,) to calculate full electrostatics for a trajectory file at
> each frame in a trajectory.
> --
> - What I have been doing so far is to set the cutoff at arbitrary large
> values (say 5000 Angs., ), which gives a small difference of ~unit
> kcal/mol for electrostatics. It's good enough, but really inelegant.
> - There is a significant difference between cutoffs and full
> electrostatics:
> Cutoff(Angs.) Electrostatic (kcal)
> 100 -239.163
> 140 -224.893
> 200 -215.19
> 5000 -204.5
> Quick script -205.38 (partially due to differences in accuracy constants
> used?)
> - We wish to compare energy calculations from NAMD simulation results with
> its components, e.g. vacuum electrostatic potential, VdW, etc.
>
>
> So, before making a customised version of the plug-in, I'm wondering if
> this has been done already?
>
> ----------------------------------------------------------------
> This message was sent using IMP, the Internet Messaging Program.
>