VMD-L Mailing List

From: David F. Green (dfgreen_at_ai.mit.edu)
Date: Thu Jul 15 2004 - 15:29:50 CDT

I would second that this seems to be a failing of psfgen to create a truly
CHARMM compatible PSF. I'm not sure exactly how psfgen works, but the
CHARMM forcefields are designed to have groups which are minimal numbers
of atoms summing to an integral charge. When PSF's are generated from
residue topology files using CHARMM, at the very least each residue is
defined as a single group. That's why the rather small maximum of 1000
atoms per group is in place .... even if a fairly large organic molecule
was contained in a single group, there would be much fewer than 1000
atoms, and all biopolymers should be defined as being composed of
individual monomers.

Perhaps the major issue is that psfgen is not designed to be used as a way
to generate PSF's for use in CHARMM, but that the "charmm-style psf" is
designed simply to have the same format that CHARMM would output. If you
really want to be doing your calculations in CHARMM, I would suggest using
the "generate" and "patch" commands to create the PSF.

+--------------------------------------------------------------------+
| David F. Green | |
| Postdoctoral Associate | Phone: 617-253-5438 |
| Division of Bioengineering and Computer | Mobile: 617-953-3922 |
| Science & Artificial Intelligence Lab | Fax: 617-258-8682 |
| Massachusetts Institute of Technology | |
| 32 Vassar St., 32-211 | E-mail: dfgreen_at_mit.edu |
| Cambridge, MA 02139 | |
+--------------------------------------------------------------------+

On Thu, 15 Jul 2004, Bogdan Costescu wrote:

> On Thu, 15 Jul 2004, Jerry Xu wrote:
>
> > ***** LEVEL -3 WARNING FROM <MAKGRP> *****
> > ***** maximum number of atoms per group exceeded
>
> OK, so this comes from the nonbonded exclusion calculations. The
> parameter that is checked against is defined as:
>
> "The maximum number of atoms in any electrostatic group"
>
> and has a value of 1000 (MAXING in source/fcm/dimens.fcm). And when
> you look earlier in the output (reformatted to avoid line wrap):
>
> Number of atoms = 1569 Number of groups = 1
>
> So you are asking CHARMM to put all your 1569 atoms in one group and
> this fails. Is this your (or psfgen's) intention ? When I did
> calculations when CHARMM I used to have an average of 3-4 atoms per
> group...
>
> --
> Bogdan Costescu
>
> IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
> Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY
> Telephone: +49 6221 54 8869, Telefax: +49 6221 54 8868
> E-mail: Bogdan.Costescu_at_IWR.Uni-Heidelberg.De
>
>
>