VMD-L Mailing List

From: Nathan C. Rockwell (ncrockwell_at_ucdavis.edu)
Date: Thu Dec 08 2005 - 15:17:48 CST

Hydrogens will be needed for worthwhile results, and the choice of hydrogen
radius is not as easy as might be thought because gromacs and other MD
packages typically use very small radii, which perform poorly in programs
like APBS or mead.

In gromacs, you should be able to add all hydrogens by using pdb2gmx and
choosing an all-atom force field such as OPLS-AA. There should also be a
pdb2pqr Python script in the APBS distribution stuff
(/tools/conversion/pdb2pqr/...) which can sometimes add the required
hydrogens and will supply correct radii in any case (ISTR the
gromacs-generated .pqr files use MD radii). Both that script and pdb2gmx
will need considerable massaging if you have ligands or similar
complications!

hope this helps,
nr

On 12/8/05 8:14 AM, "John Stone" <johns_at_ks.uiuc.edu> wrote:

> Mamoru,
> Can you send me the gromacs structure you're testing with?
> I think the hydrogens do make a difference, so if you're missing atoms,
> then before you'd use the VMD ABPS interface, you'd want to use the
> autopsf tool to build a complete PSF file first. Once loaded, _that_
> would be what you'd want to run APBS on I believe. That should also
> cure the remaining warnings you're getting. Try that and let me know
> if it works better for you. Also, if you can send me your structure
> file, I'll give it a try locally as well.
>
> John Stone
> vmd_at_ks.uiuc.edu
>
> On Wed, Dec 07, 2005 at 09:48:56PM -0600, Mamoru Yamanishi wrote:
>> Hello John,
>>
>> Thanks for your advise.
>>
>>> If you're running one of the recent test versions, I've found
>>> (and just fixed) a bug in the APBS plugin that caused it not to
>>> find the radii.dat file that VMD uses to assign radii and charge
>>> information and generate a .pqr file automatically when it runs the job.
>>> If you're running one of the recent test versions, try replacing your
>>> apbsrun.tcl script with the new one I've attached to this email.
>>> The installed apbsrun script is normally found (On unix machines) in:
>>> /usr/local/lib/vmd/plugins/noarch/tcl/apbsrun1.1/apbsrun.tcl
>>>
>>> Rename the old one to apbsrun.bak and then copy in the new one and
>>> try re-running your APBS job.
>>
>> I tried but it was the same. diff command gave followings:
>>
>> diff apbsrun.tcl apbsrun.tcl.old
>> 9c9
>> < ## $Id: apbsrun.tcl,v 1.101 2005/12/07 23:02:10 johns Exp $
>> ---
>>> ## $Id: apbsrun.tcl,v 1.100 2005/04/28 19:20:26 johns Exp $
>> 113c113
>> < set datfile [file join $::env(VMDDIR) plugins noarch tcl apbsrun1.1
>> radii.dat]
>> ---
>>> set datfile [file join $::env(VMDDIR) plugins noarch tcl apbsrun1.0
>> radii.dat]
>>
>> Do you think that permition or owner of the file/folder affect the
>> behavior?
>>
>> I have additional silly question. Using PDB2PQR server, the returned PQR
>> file
>> has lines for hydrogen atom. On the other hand, a PQR generated by
>> VMD (or APBS plugin?) has same lines as PDB file. If there is no ATOM
>> line for hydrogen, do I think both potentials are (basically) same?
>> --
>> Mamoru Yamanishi, Ph.D. <myamanishi3_at_unl.edu>
>> Research Assistant Professor
>> Dr. Banerjee's Lab., Biochemistry Department
>> University of Nebraska-Lincoln