[gmx-users] Hexagonal prisms box

[Mark Abraham]

Hi,

On Sun, Oct 4, 2015 at 10:24 AM Juan David Orjuela Zuñiga <
jd.orjuelam at uniandes.edu.co> wrote:

> Hello,
>
>
> I'm trying to simulate a membrane in a hexagonal prism (I set it up with
> INSANE Python script for Martini, as I believe is actually the option by
> default) but when I visualize it in VMD it shows as triclinic. I attempted
> a solution I found on another discussion:
>
>
> trjconv -f dopc-min-init.gro -o test.gro -ur compact -pbc cluster  -s
> dopc-min-init.tpr
>
>
> But it shows a triclinic box again, am I missing something?


The implementation of trjconv is a disaster inasmuch as if you try to do
two things in one operation, the result often depends on the order of the
operations, and it is not always clear which one does (or should) happen
first. In particular, if the implementation of -pbc cluster happens second,
then it might change the coordinates to suit that operation, so -ur doesn't
have an effect. I suggest doing trjconv -pbc cluster if you think you need
that, and then in a second call to trjconv try to set -ur compact on the
result of the first call. (And I still don't know that -ur compact will do
the right thing for a hexagonal cell, but I imagine it will, perhaps
depending on the contents of -f.) See also
http://www.gromacs.org/Documentation/Terminology/Periodic_Boundary_Conditions

Mark