Want to get MM energies (electrostatic and van der Waals) without any cut-offs

User discussions
,
1

GROMACS version:2024
GROMACS modification: No

I have a small system in which there is a ligand with several waters. I would like to simulate this and print the coulombic and LJ energies without any cutoffs. I have very small system so I do not care about the computational performance but I care about the inclusion of all solvents’ interaction with ligand. Current Gromacs version keeps asking the cut-off scheme for vdw and coul, which i really do not want. How can I achieve that? One way to increase box size but in NVT-MD solvents gets too far from ligand. In NPT-MD, the box size shrinks over time and at some point reaches to the limit of the cutoffs (twice as much). All I want is a simple MM energy term without exclusion of any solvents by cut-offs on vdw or coul.
So far what I had to do is the following:

  1. I bring the system at the equilibrium at NPT after NVT. That way, I could learn how much the box size would be according to the number of solvents I have in the system (which is less than 100).
  2. I use gmx editconf to change the box size of the NPT to trick the program for accounting all interactions
  3. I rerun mdrun -rerun of the NPT trajectory.

Is this an accurate way? or could someone tell me how to achieve this step by step?

2

I have found the solution using Gromacs <=2020:

  1. Run the simulation regularly using either group or verlet scheme. (EM, NVT and NPT)
  2. Convert the trajectory of interest by:
    “gmx trjconv -f nptsd.xtc -s nptsd.tpr -pbc mol -center -o trjmol.xtc”
  3. Generate another tpr by a special mdp in which pbc=no, nstlist=rlist=rcoulomb==rvdw=0 and cutoff-scheme=group
    “gmx grompp -f test.mdp -p topol.top -c npt.gro -o test.tpr”
  4. Re-run the trajectory by:
    “gmx mdrun -rerun trjmol.xtc -deffnm test”

Now the new edr file has the correct coulomb and vdw, agreeing with gmx_MMPBSA or g_MMPBSA predicted coul and vdw terms.

1 Like