[gmx-users] Shrink simulation box while maintaining ensemble

[Justin Lemkul]

On 2/15/16 10:35 AM, Matthias Ernst wrote:
> Hi,
>
> I am running a protein folding simulation, starting with an extended
> structure that needs a lot of water molecules to fulfill the minimial
> image convenction.
> Is there a way to "shrink" a simulation box and get rid of unnecessary
> solvent molecules while keeping velocities (to preserve the ensemble as
> far as possible and avoid re-equilibration)?
>

You could define a group of atoms that satisfies some geometric criterion 
(boundaries for x,y,z) with gmx select, and use trjconv with that index group to 
save only those atoms.  Assuming you're doing this on a .trr file with 
velocities, those  would be preserved.  But the complication is that you then 
have a new box (which you have to define), which puts particles at new positions 
relative to the new boundaries, and those velocities and positions are not 
necessarily compatible with this new definition (you could have atoms at the new 
boundaries that are about to crash right into one another!), so you probably 
have to re-equilibrate, anyway.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================