[gmx-users] membedded.gro from g_membed has broken molecules (fix periodicity?)

[Mark Abraham]

On 8/04/2011 1:58 PM, Peter C. Lai wrote:
> The output of g_membed is membedded.gro which appears to have broken lipids
> at the box boundaries, making the box "smooth" (as if g_membed did not fix
> the periodicity of the final frame) when visualized in VMD.

Yep. g_membed is a bastardised form of mdrun, which has the same properties.

>   How do I fix the
> periodicity for visualization - I cannot use trjconv -s membed.tpr because
> it complains about the missing atoms from the original tpr, as g_membed
> deletes molecules on the fly after calculating the overlaps...

You can use tpbconv to select a subset of atoms in .tpr (to make a 
non-fully-functional .tpr), or equivalently trjconv to select a subset 
of the initial configuration in the structure file that created that 
.tpr. Those will have whole molecules.

> Perhaps more importantly, should I try to make the membedded.gro whole
> before using it as the input coordinate file for the subsequent equilibration
> md? The g_membed paper does not mention this...

I don't think you need to - if you your concern is whether mdrun will 
cope with the broken molecules.

Mark