On Mon, Jan 6, 2014 at 3:28 PM, Brian Radak <radak004.umn.edu> wrote:
> While preparing trajectories for post processing via MM/GBSA, I used the
> following procedure:
>
> Using cpptraj:
> 1) center/image an explicit solvent trajectory
> 2) strip out the solvent
> 3) save a "stripped" trajectory
>
> and then one of the following:
>
> Using sander:
> 1) use imin = 5 for some energy calculations --> This runs without error
>
> Using MMPBSA.py
> 1) input the stripped prmtop and mdcrd file --> This fails because IFBOX =
> 2.
>
> I got around this by changing the prmtop by hand (sorry, I no longer have
> the output error message from the mdout file), which was pretty simple.
>
> I don't know what, if any, part of this is a bug. Presumably a stripped
> trajectory should always have IFBOX = 0? But that doesn't seem strictly
> necessary. By the same token, why does MMPBSAY.py require IFBOX = 0 with
> igb > 0, if that information is just ignored by sander?
>
ifbox should only be set to non-zero values when "solvateOct",
"solvateBox", or "setBox" is used in tleap, which means that solvent was
explicitly set. Only recently have tools other than tleap been able to
modify topology files and write out new ones. MMPBSA.py is strict because
it catches common errors (if someone uses a solvated topology file for what
should be a stripped topology file). 99% of the time (and it used to be
100% of the time), a IFBOX>0 topology file should never be used with
MMPBSA.py as either the complex, receptor, or ligand topology files.
ante-MMPBSA.py (and ParmEd), sets ifbox to 0 if all of the solvent is
stripped, while it seems that cpptraj does not. I would say there's no
real 'bug' here, but a difference between expected and documented behavior
of cpptraj and MMPBSA.py in this case. You can see if "nobox" works to get
rid of the box information in cpptraj (in which case, it's no bug, just
somewhat inconsistent behavior b/w amber's different programs).
Good luck,
Jason
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Jan 06 2014 - 13:00:04 PST
