[gmx-users] Odd behaviour of editconf or VMD?

[Jernej Zidar]

Hi,
  I am trying to solvate a small DNA molecule. Since I don't want so
simulate more waters than what is really neccessary, I decided to try the
dodecahdron and octahedron unit cells.

  After I added waters I went to examine the resulting structures in VMD
only to find that the octahedron and dodecahedron cells are completely
wrong. Instead of the familiar shape (roughly a cube with trimmed edges for
dodeachedron), I am greeted with a parallelepiped-like shape.

  I tried Gromacs 5.0.7, 5.1.3, 2016 but the results seem to be always the
same. Is this an expected behaviour? How do I "reconstitute" the resulting
structure?

Thanks,
Jernej