[gmx-users] Difference between output_coordinates.gro and trajectory.xtc coordinates

[Mark Abraham]

Hi,

On Thu, Jun 22, 2017 at 11:05 AM Diez Fernandez, Amanda <
amanda.diez10 at imperial.ac.uk> wrote:

> Hi Mark,
> Thank you for your reply.
>
> It is exactly the opposite though:


OK, I understood the images the other way around, without titles.


> for output_coordinates.gro which is the
> output from:
>
>  mdrun Š -c output_coordinates.gro
>
> it seems like atoms are diffusing out of the box.
>

They're just in whichever representation was convenient for mdrun, which as
I said is not intended to be the same as the output of any of the things
you might do with trjconv.

In the trajectory, I allow atoms to be seen to diffuse out of the box,
> using -pbc nojump, and I don¹t see them diffuse out (They stay within the
> unit cell, which is what I expect being a silica substrate at room
> temperature over a very short period of time).
>

Sure.


> So what puzzles me is why for the output of (mdrun -c) the atoms seem to
> diffuse out, specially so if I cannot see this in the trajectory.
>

mdrun will generally "make molecules whole" for -c, but otherwise doesn't
care. Implementing general triclininc 3D periodicity with domain
decomposition is a messy business. If you've written the final frame to the
trajectory, you can use trjconv to make the two outputs have the same
representation, but they still might be different by a translation.

Mark

Sorry if I didn¹t explain it very clearly last time.
>
> Thanks!
>
>