[gmx-users] energy groups rerun from reduced xtc file

[Harry Mark Greenblatt]

BS"D

Dear Mark,

    Thank you for the advice; removing the checkpoint file does allow grompp to produce a new tpr file for mdrun.

In the case, however, where the energy groups were defined before the run, and so it should be a simple
case of rerunning mdrun on CPU's, the use of convert-tpr to produce a smaller tpr file gives the following error in
mdrun:

Software inconsistency error:
Position restraint coordinates are missing

(the DNA is restrained in these runs).

So producing a new tpr file via grompp seems the only way to  go...?

Thanks

Harry




Harry M. Greenblatt
Associate Staff Scientist
Dept of Structural Biology           harry.greenblatt at weizmann.ac.il
Weizmann Institute of Science        Phone:  972-8-934-3625
234 Herzl St.                        Facsimile:   972-8-934-4159
Rehovot, 76100
Israel

________________________________________
From: gromacs.org_gmx-users-bounces at maillist.sys.kth.se [gromacs.org_gmx-users-bounces at maillist.sys.kth.se] on behalf of Mark Abraham [mark.j.abraham at gmail.com]
Sent: Thursday, December 04, 2014 1:32 PM
To: Discussion list for GROMACS users
Subject: Re: [gmx-users] energy groups rerun from reduced xtc file

On Thu, Dec 4, 2014 at 10:16 AM, Harry Mark Greenblatt <
harry.greenblatt at weizmann.ac.il> wrote:

> BS"D
>
> Dear All,
>
>    In an attempt to save disk space while using longer simulations of a
> DNA + Protein + solvent system,
> I had mdrun create an xtc file with only the DNA, Protein, and ions,
> excluding the water molecules.
> No trr file was produced.  This is fine for many things, but since I also
> want to look at energies between
> defined groups, and the GPU-based calculations do not support this, I need
> to to a "rerun" of the simulation with CPU's only,
> to do the energy calculations.
>
> If I understood correctly, the procedure to change or introduce energy
> groups would be:
>
> grompp -f yournew.mdp -c yourold.gro -t yourold.cpt -p your.top -o
> newrerun.tpr -n yournewindexfile.ndx
>

There's no need to provide the checkpoint file. The rerun takes the
coordinates of the old trajectory and applies the model physics in the
.tpr. It doesn't generate new coordinates, so questions of
trajectory/ensemble continuity do not arise. Thus, no checkpoint needed.


> mdrun -v -s newrerun.tpr -rerun yourold.xtc -e new.edr
>
> The problem is that the xtc file and the initial topology no longer have
> the same number of atoms.  I have gone through a lot of conversions,
> to create new top and gro files without water molecules.
>

grep -v SOL does most of the work for coordinate files, and the topology
change should be about one line in [system].

Alternatively, gmx convert-tpr can make subsets of .tpr files, which should
also be enough for your purposes.

I was not, however, able to take care of cpt file, since my attempt to use
> trjconv succeeded, but grompp does not like it because trjconv
> does not output a cpt file.
>
> Is there a solution to this problem (other than having to produce all-atom
> xtc or trr files)?
>

Yep, don't do work you don't need :-)

Mark