VMD-L Mailing List

From: Axel Kohlmeyer (akohlmey_at_gmail.com)
Date: Sun Feb 07 2010 - 14:05:19 CST

On Sun, 2010-02-07 at 12:51 -0500, Parisa Akhski wrote:
>
>
> Dear all,
>
dear parisa,

> I am trying to generate a psf file for a two layer molecule which
> consists of four guanine molecules in each layer. There is also a
> sodium ion between the layers. If I use the pdb file which is using
> MOL after the name of each atom and use AutoPSf to generate the psf
> file, I come up with the message"Autopsf has detected 1 unparametrized
> component" with the (1:MOL 4atoms)in a box specifying the atom. If I
> use the GUA instead of MOL as I am dealing with guanine molecules it
> tells me that the psf file has been completed but, neither VMD nor
> Rasmol could open the file. Also the total number of 129 atoms in the

unfortunately, you will first have to understand that autopsf was
written primarily to make the process or converting "standard" files
into topologies. yours is non-standard in a few ways. so you first
have to learn and understand how the psfgen topology builder works,
before you can be successful. the most important fact to realize is
that no simple computer program understands chemistry per se. all
they can do is either guess or look up the result in some database,
provided the input follow some conventions.

psfgen does the latter. i first tries to match your residue names
(MOL GUA) with names in its topology database (the charmm topology
file that you have to specify or the one that it uses by default).
if _you_ know that you have guanidine, then you have to use the
exact proper residue name for it. using MOL is thus nonsense.
secondly, it will try to match the given atom names with the atom
names in the residue template, so that it can assign the force field
atom types. your atoms all only have the element names, so that is
not going to work. hence the messages about non-recognized atoms.

finally, you have multiple copies of the same molecule. in that case
it is often better to split the file into multiple segments and then
try to build a topology for only one of them. that topology is
independent from the coordinates and can thus be reused. you can
combine multiple segments easily with psfgen (this is how solvate
and the membrane builder work).

> original pdb file reduces to 33 atoms in autopsf file. I tried
> different names for sodium atom(GUA, HETATM and even without the
> sodium atom) but, all give me the same result. I have also used a
> different DNA pdb file directly from pdb.org(pdb ID=325d) to make the
> psf file but, I see the error"autopsf has detected 88 unparametrized
> components". I am just wondering whether anyone has any experience in
> this case or not. I have attached the guanine pdb file with the name
> of MOL for all atoms.

when downloading data from the PDB, effectively the same restrictions
apply. those files usually follow some naming conventions, but they
are not always the same as in the charmm topology files, or sometimes
it contains residues that are not (yet) parameterized in the topology
file that you are using.

i recall that this kind of question has come up several times in the
past, so i suggest you search the mailing list archives and also have
a look at the psfgen tutorial. you are obviously dealing with a system,
where the "lets just click on some buttons and hope for the best"
approach has very little chances to be successful. no pain, no gain.

good luck,
    axel.

> Thanks a lot
>
> Parisa
>
>
>
>
> 

-- 
Dr. Axel Kohlmeyer  akohlmey_at_gmail.com 
Institute for Computational Molecular Science
College of Science and Technology
Temple University, Philadelphia PA, USA.