From: Francesco Pietra (
Date: Fri Jun 12 2015 - 11:13:50 CDT

I forgot to mention that the full recomposition of the system holds for the
last frame 4999 only. Should the "painfully slow unwrap process" be used in
order to run the movie of the simulation?

---------- Forwarded message ----------
From: Francesco Pietra <>
Date: Fri, Jun 12, 2015 at 5:53 PM
Subject: Re: vmd-l: unwap/wrap failure in recomposing splitted system from
periodic cells
To: Josh Vermaas <>
Cc: VMD Mailing List <>

Hi Josh:

I tried again, hitting gold.

Looking better at the end of the simulation, I noticed that the tetramer
had now been separated into two dimers. On one side of the box, PRD half
out the box and PRB entirely within the box. At the opposite side, PRA
nearly totally out of the box and PRC entirely within the box.

Repeating your commands with *segname PRC", I got the whole system
perfectly in orde, including all ligands and potential ligands, within the
same box.

I can safely rule out errors from my side in my previous failed attempts.

No words enough to thank you.

PS: Do you have any idea whether that could be extended to bigdcd in order
to follow certain specific rgyr along the whole trajectory? (with 64GB at
the computing center I am limited to load three 24h simulations)

On Fri, Jun 12, 2015 at 4:47 PM, Josh Vermaas <> wrote:

> Hmm... In principle it shoulda totally worked, but you can try things
> like removing the -sel all argument? Admittedly I was going on memory, so
> maybe there is something weird about all not being in quotes. I like how it
> complains about an unknown option after having used it the previous command.
> -Josh
> On 6/12/15 2:33 AM, Francesco Pietra wrote:
> Hi Josh:
> With "segname PRA" in place of your "segname A", for the same purpose,
> Main console display active (Tcl8.5.6 / Tk8.5.6)
> (MD_O2-out) 1% pbc wrap -sel "protein" -centersel "segname PRA" -center
> com -compound fragment
> Info) 100.0% complete (frame 4999)
> >main< (MD_O2-out) 2 % pbc wrap -sel all -centersel "protein" -center com
> -compound fragment
> error; pbcwrap: unknown option: -compound
> >Main< (Md_O2-out) 3 %
> Is that understandable why error?
> Thanks a lot
> francesco
> On Thu, Jun 11, 2015 at 9:31 PM, Josh Vermaas <>
> wrote:
>> Here is how I do it on my own systems, which avoid the painfully slow
>> unwrap process:
>> pbc wrap -sel "protein" -centersel "segname A" -center com -compound
>> fragment
>> pbc wrap -sel all -centersel "protein" -center com -compound fragment
>> The trick here is that I have one monomer (segname A) that I use to bring
>> the other potentially split monomers into the same pbc box. The unwrap
>> command followed by a wrap would also work, but you forgot the "compound
>> fragment" argument, so the algorithm did what you told it to, and that is
>> to wrap atoms around into a rectangle regardless of connectivity, resulting
>> in ludicrously long bonds.
>> -Josh Vermaas
>> On 6/11/15 12:24 PM, Francesco Pietra wrote:
>> Hello:
>> Following the splitting of a homotetramer into intact trimer/monomer at
>> ca 300 ns trajectory, I tried the following three commands in the tk
>> console of the remote visualization for the last 24h simulation (ca 11 GB
>> file size). I had 64GB memory available (the max I could have)
>> set all [atomselect top all]
>> pbc unwrap -sel all (which operated very slowly on the 4999 frames)
>> pbc wrap -sel all
>> The latter created a box of lines of the original size. Then, command
>> "all not water" removed those pertaining to water, leaving lines not
>> amenable to any "Drawing Method". These lines looked like long bonds,
>> which were not present in the splitted system.
>> I used "all" as the system is also composed of ligands, some firmly
>> residing into binding pockets, some other ones traveling from the
>> surrounding medium to their binding pockets.
>> The "intact" in the first line above is not entirely correct. That is,
>> the monomer is removed from the homotetramer without its major ligand (the
>> substrate to be modified by the enzyme).
>> Thanks for suggestions about how to amend my faulty commands.
>> francesco pietra