From: JC Gumbart (
Date: Tue Aug 09 2011 - 23:52:22 CDT

Sent off list

-----Original Message-----
From: [] On Behalf Of
John Stone
Sent: Tuesday, August 09, 2011 10:19 PM
To: JC Gumbart
Cc: Anna Whitney; vmd-l list
Subject: Re: vmd-l: Long atom types from CGenFF causing charge problems

  If someone can provide me with an example PSF file that exhibits
this problem, and it is a legal PSF file, it should be easy to modify
the PSF plugin to behave correctly. The PSF plugin already handles
PSF files that use EXT and other flags, so unless these files are either
using another new variant of the PSF format or they are illegally formatted,
I don't know why they would not read correctly.


On Tue, Aug 09, 2011 at 05:42:32PM -0500, JC Gumbart wrote:
> If you calculate the ions needed to neutralize and reach a given
concentration yourself, you can use autoionize to add them with the -nions
flag. (Note that concentration = #ions * 55.5/ # water molecules ). NAMD
will sum the charge correctly and confirm your system is neutral.
> At the risk of sounding demanding, this is something that has been raised
many times by myself and others for months and REALLY needs to be fixed if
VMD is going to be usable for modern forcefields.
> On Aug 9, 2011, at 4:01 PM, Anna Whitney wrote:
> > Hi,
> >
> > I'm trying to add ions to a system consisting of a protein with a ligand
parameterized using CGenFF, a DOPC/DOPE membrane, and water. VMD's
Autoionize plugin returns a non-integral charge, however. When I check the
charge on the ligand using
> > set sel [atomselect top "resname TAK"]
> > measure sumweights $sel weight charge
> > I also get a non-integer (and a much higher charge than I should), while
all other components of the system have integral charges according to VMD. I
don't think Autoionize is giving me anything useful, based on how wrong it
is about the initial charges.
> >
> > Previous messages on the mailing list seem to indicate that this problem
can be caused by VMD misreading PSF files with atom types longer than four
characters, which my PSF does have, from CGenFF. However, I couldn't find a
workaround on the mailing list or anywhere else- is there any way to get VMD
to correctly read charges from a PSF that includes
longer-than-four-character atom types? Alternately, is there some other way
of adding ions (to neutralize and/or bring to a certain concentration) that
doesn't involve VMD's interpretation of the charge from the PSF?
> >
> > Thanks,
> > Anna Whitney

NIH Resource for Macromolecular Modeling and Bioinformatics
Beckman Institute for Advanced Science and Technology
University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801           Phone: 217-244-3349       Fax: 217-244-6078