From: Josh Vermaas (vermaas2_at_illinois.edu)
Date: Wed Jun 24 2015 - 16:46:52 CDT

Hi Marawan,

In cases where the protein gets split, you have two options. The
canonical approach is an unwrap followed by a wrap, and the other is to
wrap twice using different arguments for centersel. What is going on
when you say "centersel protein" is that pbctools will calculate the
center of the protein atomselection as it currently stands. Since the
center of your split protein isn't where you want it to be, it won't do
what you expect. Unwrap will remove "jumps" where an atom moves further
than half a periodic box length, and then a subsequent wrap command will
work. The other option is to wrap twice, once with something like
"centersel segname A", which will put the protein back together again,
and one more time with "centersel protein" so that it is centered in the
membrane.

-Josh Vermaas

On 6/24/15 3:51 PM, Marawan Hussien wrote:
> Hi,
> I am simulation an ion channel embedded within a POPC membrane built
> through CHARMGUI and simulated with NAMD . The ion channel assembled
> from 3 domains and are not covalently connected. The problem is that
> when i visualize the trajectories i observed that the protein domains
> are translated through the X axis and due to PBC, some domains appear
> from the other side of the membrane. I tried to use the PBCtools to
> reorient the protein to the center of the visualization cell but in vain.
> I tried this command:
> pbc wrap -center origin -compound residue -all
>
> And:
>
> pbc wrap -centersel "protein" -center origin -compound residue -all
>
> But none of them work even when I read the XST file using the readxst
> command
>
> Any suggestions?
>
> Regards,
> Marawan