VMD-L Mailing List

From: Jim Phillips (jim_at_ks.uiuc.edu)
Date: Thu Jun 09 2005 - 13:41:38 CDT

Hi,

The whole point of psfgen was to match coordinates in random PDB files
with the atoms in the residue definition in the topology file.
Therefore, the coordinate reading code in psfgen uses segment, residue,
and atom names, and it expects these to be unique. Clearly this isn't the
case with the current implementation of LES. The solution is to run
multiply last, after the structure has been completely assembled.

Another option, if that isn't possible, is to use VMD atom selections to
copy x, y, z, beta, and occupancy for the segments with replicated atoms
from the old pdb to the new, something like this:

set orig_mol [mol load pdb original.pdb]
set solv_mol [mol load pdb solvated.pdb]
set orig_sel [atomselect $orig_mol "segname IH01"]
set solv_sel [atomselect $solv_mol "segname IH01"]
$solv_sel set {x y z occupancy beta} \
                  [$orig_sel get {x y z occupancy beta}]

-Jim

On Thu, 9 Jun 2005, Marcin Wojciechowski wrote:

> Hello to All,
> Dring preparation of the system for MD I encountered the following
> problem:
>
> I created initial psf and pdb file of the system using psfgen with 5
> copies of the ligand needed for Locally Enhanced Sampling MD in NAMD.
>
> The fragment of my pdb:
>
> ATOM 1 C18 C130 326 3.254 7.549 18.595 1.00 1.00 IH01
> ATOM 2 C18 C130 326 3.254 7.549 18.595 1.00 2.00 IH01
> ATOM 3 C18 C130 326 3.254 7.549 18.595 1.00 3.00 IH01
> ATOM 4 C18 C130 326 3.254 7.549 18.595 1.00 4.00 IH01
> ATOM 5 C18 C130 326 3.254 7.549 18.595 1.00 5.00 IH01
> ATOM 6 C1 C130 326 2.969 8.913 18.728 1.00 1.00 IH01
> ATOM 7 C1 C130 326 2.969 8.913 18.728 1.00 2.00 IH01
> ATOM 8 C1 C130 326 2.969 8.913 18.728 1.00 3.00 IH01
> ATOM 9 C1 C130 326 2.969 8.913 18.728 1.00 4.00 IH01
> ATOM 10 C1 C130 326 2.969 8.913 18.728 1.00 5.00 IH01
> etc.
>
> every atom specified for LES is multiplied 5 times here and the BETA
> column designates the number of copy (ant that's ok). However when I
> solvate such a pdb using simple "solvate system.psf system.pdb -t 15 -o
> system.solv" I get the following output from system.solv.pdb:
>
> ATOM 1 C18 C130 326 3.254 7.549 18.595 1.00 0.00 IH01
> ATOM 2 C18 C130 326 0.000 0.000 0.000 -1.00 0.00 IH01
> ATOM 3 C18 C130 326 0.000 0.000 0.000 -1.00 0.00 IH01
> ATOM 4 C18 C130 326 0.000 0.000 0.000 -1.00 0.00 IH01
> ATOM 5 C18 C130 326 0.000 0.000 0.000 -1.00 0.00 IH01
> ATOM 6 C1 C130 326 2.969 8.913 18.728 1.00 0.00 IH01
> ATOM 7 C1 C130 326 0.000 0.000 0.000 -1.00 0.00 IH01
> ATOM 8 C1 C130 326 0.000 0.000 0.000 -1.00 0.00 IH01
> ATOM 9 C1 C130 326 0.000 0.000 0.000 -1.00 0.00 IH01
> ATOM 10 C1 C130 326 0.000 0.000 0.000 -1.00 0.00 IH01
> etc.
>
> Here the coords of the copies are reseted to 0.0 also BETA "counters"
> and occupancies are set to invalid value.
>
> Is solvate module incompatible with pdb produced with psfgen with LES
> copies of the atoms? Anyone encountered similar problem? Any
> suggestions?
>
> Best regards, Marcin
>
> --
>
> ----------------------------------------
> Marcin Wojciechowski, MSc
>
> Laboratory of Molecular Modeling
> Department of Pharmaceutical Technology
> and Biochemistry
> Gdansk University of Technology
> Narutowicza St 11/12
> 80-952 Gdansk
> POLAND
>
> tel 4858-3471596
> fax 4858-3471144
>
> nitramus_at_hypnos.chem.pg.gda.pl
> www.pg.gda.pl/chem/katedry/leki/leki.htm
> ----------------------------------------
>