[gmx-users] Squishing or Stretching Membranes

[Mirco Wahab]

On 09.07.2013 16:11, Neha wrote:
> I had a question about trjconv. After one of my simulations has ended I want
> to use the final structure file to run some other simulations. However, what
> I want to do is run an NVT run using the average box size of the earlier
> run. Since the final structure file will most likely not be at the exact
> average I was wondering if I could use trjconv -pbc mol to put all the atoms
> in a box either smaller or bigger than the original.

You could dump many structures from the last part of the trajectory, and:
  - take one structure that has all box vectors almost at, but *below*
    your target size,
  - edit the bottom line of the coordinate file to the desired size,
  - run some steps steepest descent minimization in order to correct
    for some overstretched bonds of molecules crossing the PBC,
  - run

The absolute last configuration of your run has no "specific meaning"
in a statistical sense. You can use any other configuration from
the equilibrium-part of the run.

my € 0.02