[gmx-users] persistent jumps in trajectories

[Justin Lemkul]

On 5/4/16 6:12 AM, morpheus wrote:
> Hi,
>
> I'm having problems getting jumps out of my trajectories. Usually I follow
> the suggestion of the Gromacs website (
> http://www.gromacs.org/Documentation/Terminology/Periodic_Boundary_Conditions)
> and convert my trajectories with the perl script below: remove water, make
> it whole, cluster, nojump, center and fit.
>
> For previous projects these steps always did the job right.
>
> But unfortunately this does not work for long simulations (1500 ns) of
> PDB-ID 2RVD (a peptide of only 10 amino acids). I have performed the
> simulation multiple times and after about 300-500 ns the jumps  still start
> even after processing the trajectories as given below. With jumps I mean
> this: http://s32.postimg.org/z8p90wf51/jump.jpg
>
> Doing analysis on the trajectories with e.g. g_dist yields “There were 4
> inconsistent shifts. Check your topology”.
>
> I am wondering what I am doing wrong and how can I fix the trajectories?
>
>
> Here is the script that I am usually using:
>
>
> my $bn = $ARGV[0]; # base name
>
> $cmd = $execPath.$gPrefix."trjconv".$gPostFix." -f ".$bn.".md.xtc -s
> ".$bn.".md.tpr -o ".$bn.".noWater.md.xtc -skip 10 << Ende \n1\n1\n << Ende";
>
> system $cmd;
>
>
> $cmd = $execPath.$gPrefix."trjconv".$gPostFix." -f ".$bn.".noWater.md.xtc
> -s ".$bn.".md.tpr -o ".$bn.".noWater.whole.md.xtc -pbc whole << Ende
> \n1\n1\n << Ende";
>
> system $cmd;
>
>
> $cmd = $execPath.$gPrefix."trjconv".$gPostFix." -f
> ".$bn.".noWater.whole.md.xtc -s ".$bn.".md.tpr -o
> ".$bn.".noWater.whole.cluster.md.xtc -pbc cluster << Ende \n1\n1\n << Ende";
>
> system $cmd;
>
>
> $cmd = $execPath.$gPrefix."trjconv".$gPostFix." -f
> ".$bn.".noWater.whole.cluster.md.xtc -s ".$bn.".md.tpr -o
> ".$bn.".noWater.whole.cluster.nojump.md.xtc -pbc nojump << Ende \n1\n1\n <<
> Ende";
>
> system $cmd;
>
>
> $cmd = $execPath.$gPrefix."trjconv".$gPostFix." -f
> ".$bn.".noWater.whole.cluster.nojump.md.xtc -s ".$bn.".md.tpr -o
> ".$bn.".noWater.whole.cluster.nojump.center.md.xtc -center << Ende \n1\n1\n
> << Ende";
>
> system $cmd;
>
>
> $cmd = $execPath.$gPrefix."trjconv".$gPostFix." -f
> ".$bn.".noWater.whole.cluster.nojump.center.md.xtc -s ".$bn.".md.tpr -o
> ".$bn.".final.md.xtc -fit rot+trans << Ende \n1\n1\n << Ende";
>
> system $cmd;
>
>
> Can anyone help?
>

For something as simple as a 10-aa peptide, all you need is:

trjconv -pbc mol -center

Center on the protein and strip the solvent if you like by just selecting 
protein as the output.  Fit afterwards if you need to do so.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================