Re: Some water molecules are missing in psf

From: Vermaas, Joshua (Joshua.Vermaas_at_nrel.gov)
Date: Thu Nov 02 2017 - 18:13:28 CDT

Hi Amali,

How did you try to generate a psf file? There are about a bazillion
different ways of doing it, and the workaround depends on how you do it.
What I *think* is going on is that all of your 96000 waters are trying
to be assigned to the same segment. pdbs only have 4 characters for the
resid, so those numbers are between 0-9999 in the original file. psfgen
then comes along and reads your water segment from the pdb you provided.
It will dutifully read in each water molecule, but since some of the
resids repeat, it will just overwrite the positions of the old water
with that resid, rather than making a new water molecule, giving you the
symptoms you see there. What I've found to be the easiest workaround for
me is to just strip the water and add it back in.

#Add an appropriate topology command here
resetpsf
set watsel [atomselect top "water"]
set notwater [atomselect top "not water"]
$watsel writepdb water.pdb
set watnum [llength [lsort -unique -integer [$watsel get residue]]]
segment WA {
for { set i 1 } { $i <= $watnum } { incr i } {
residue $i TIP3
}
}
writepsf water.psf

mol load psf water.psf pdb water.pdb
package require topotools
set mol [::TopoTools::selections2mol [list $notwater [atomselect top
"all"]]]
set mergesel [atomselect $mol "all"]
$mergesel writepsf merged.psf
$mergesel writepdb merged.pdb

Basically what I'm doing here is saving the water section alone, writing
a clean PSF for it, and then merging the water and not water pieces of
the system into a new molecule.

-Josh

On 11/02/2017 04:35 PM, Amali Guruge wrote:
> Dear All,
>
> I want to simulate a system which contains nearly 96000 water
> molecules and 700 PEG molecules. I modeled the system and I got top
> and parm files from ParaChem. When I tried go generate psf file, it
> generates a pdb with only 9999 water molecules. What could be the
> reason for this? Appreciate, if anyone can help me.
>
> Thank you.

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2018 - 23:20:40 CST