CHARMM-GUI script results in consecutive trajectories with the same start times

User discussions
1

GROMACS version: 2019.3
GROMACS modification: No

Hi all,

I’m using the run script that CHARMM-GUI provides, which involves running 10 production simulations one after the other. The relevant grompp command is as follows:

gmx grompp -f ${prod_prefix}.mdp -o ${istep}.tpr -c ${pstep}.gro -t ${pstep}.cpt -p topol.top -n index.ndx

Here, ${pstep} sets the name of the output of the previous simulation, and ${istep} sets the name of the new (current) simulation. Please see the very bottom of this post for the full submission script. I would expect that each time, the new simulation continues where the previous one left off. However, when I run trjcat on the resulting .trr files:

gmx trjcat -f step7*trr -o step7_all.trr

I see the following:

Summary of files and start times used:

          File                Start time       Time step
---------------------------------------------------------
             step7_10.trr        0.000 ps      100.000 ps
              step7_1.trr        0.000 ps      100.000 ps WARNING: same Start time as previous
              step7_2.trr        0.000 ps      100.000 ps WARNING: same Start time as previous
              step7_3.trr        0.000 ps      100.000 ps WARNING: same Start time as previous
              step7_4.trr        0.000 ps      100.000 ps WARNING: same Start time as previous
              step7_5.trr        0.000 ps      100.000 ps WARNING: same Start time as previous
              step7_6.trr        0.000 ps      100.000 ps WARNING: same Start time as previous
              step7_7.trr        0.000 ps      100.000 ps WARNING: same Start time as previous
              step7_8.trr        0.000 ps      100.000 ps WARNING: same Start time as previous
              step7_9.trr        0.000 ps      100.000 ps WARNING: same Start time as previous
            step7_all.trr        0.000 ps      100.000 ps WARNING: same Start time as previous

Why are the starting times all 0 ps? What am I missing?

#!/bin/csh
#SBATCH --job-name=membrane3W
#SBATCH --output=out.dat
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=1
#SBATCH --gres=gpu:1
#SBATCH -c 32
#SBATCH -p gpu
#SBATCH --time=2-00:00:00

set init = step5_input
set rest_prefix = step5_input
set mini_prefix = step6.0_minimization
set equi_prefix = step6.%d_equilibration
set prod_prefix = step7_production
set prod_step   = step7

# minimization
module load GROMACS/2019.3-foss-2018a
gmx grompp -f ${mini_prefix}.mdp -o ${mini_prefix}.tpr -c ${init}.gro -r ${rest_prefix}.gro -p topol.top -n index.ndx
module unload GROMACS/2019.3-foss-2018a
module load GROMACS/2019.3-foss-2018b-CUDA-10.0.130
gmx mdrun -v -deffnm ${mini_prefix}

# equilibrattion
set cnt = 1
set cntmax = 6

while ( ${cnt} <= ${cntmax} )
    @ pcnt = ${cnt} - 1
    set istep = `printf ${equi_prefix} ${cnt}`
    set pstep = `printf ${equi_prefix} ${pcnt}`
    if ( ${cnt} == 1 ) set pstep = ${mini_prefix}

    module unload GROMACS/2019.3-foss-2018b-CUDA-10.0.130
    module load GROMACS/2019.3-foss-2018a

    gmx grompp -f ${istep}.mdp -o ${istep}.tpr -c ${pstep}.gro -r ${rest_prefix}.gro -p topol.top -n index.ndx

    module unload GROMACS/2019.3-foss-2018a
    module load GROMACS/2019.3-foss-2018b-CUDA-10.0.130
    gmx mdrun -ntmpi 1 -ntomp 32 -v -deffnm ${istep}
    @ cnt += 1
end

# production
set cnt    = 1
set cntmax = 10

while ( ${cnt} <= ${cntmax} )
    @ pcnt = ${cnt} - 1

    module unload GROMACS/2019.3-foss-2018b-CUDA-10.0.130
    module load GROMACS/2019.3-foss-2018a
    set istep = ${prod_step}_${cnt}
    set pstep = ${prod_step}_${pcnt}

        if ( ${cnt} == 1 ) then
        set pstep = `printf ${equi_prefix} 6`
        gmx grompp -f ${prod_prefix}.mdp -o ${istep}.tpr -c ${pstep}.gro -p topol.top -n index.ndx
        else
	gmx grompp -f ${prod_prefix}.mdp -o ${istep}.tpr -c ${pstep}.gro -t ${pstep}.cpt -p topol.top -n index.ndx
        endif

    module unload GROMACS/2019.3-foss-2018a
    module load GROMACS/2019.3-foss-2018b-CUDA-10.0.130
    gmx mdrun -ntmpi 1 -ntomp 32 -v -deffnm ${istep}
    @ cnt += 1
end
2

This is the expected outcome when one uses grompp to generate new inputs without changing the tinit variable. It’s a bit of an outdated approach, but you can overcome it with the -settime option of trjcat.

3

Thank you so much for your response. As far as I can tell, tinit is an .mdp file option, and it can’t be changed while running grompp, right? I’m asking because the current script is set up in such a way that it uses the same .mdp file to generate consecutive .tpr files. In other words, to use the tinit option, I would have to create a different .mdp file for each run, right?

That being said, I suppose I could just run a single, longer production run and not deal with this at all.

4

This would be a better approach. The CHARMM-GUI scripts are robust and work well but they rely on old conventions (because they are unlikely to change from version to version). There’s no requirement to use them, though they are a good reference for proper implementation.