Re: [AMBER] solvatebox

From: M. L. Dodson <>
Date: Sat, 06 Nov 2010 22:56:58 -0500

On 11/6/2010 9:22 PM, Bill Ross wrote:
>> solvatebox resname MEOHBOX 0
> I think if you want resname to have a box with 10A of repeated MEOHBOX,
> you would
>> solvatebox resname MEOHBOX 10
> and this is how you would adjust the size of the box.
>> help solvatebox
> may be useful.
> Bill


In addition to the other comments, there will be very high initial VDW
energies since the solvent boxes were not constructed to be placed
immediately adjacent with no space in between. I expect the energies
may be high enough to be NaN's until you move the boundary solvent
molecules just enough for the boxes to fit closely together. You may
be able to solve this problem with a series of sequential very short
minimizations. 6-10 minimization may be enough based on my experience
with packmol-generated systems, but it may also not be possible to
overcome the VDW clashes. (The symptoms are sander crashing with
segmentation faults.) I expect the less space between the boxes, the
larger the number of minimizations that will be required. As I
learned it, the canonical AMBER way is to solvate with space between
the boxes (?leap will do this automatically), then equilibrate the
density with NPT dynamics. This is much, much easier unless you
absolutely cannot start out this way.

Bud Dodson

M. L. Dodson
Business Email: activesitedynamics-at-comcast-dot-net
Personal Email: mldodson-at-comcast-dot-net
Mobile Phone: eight_three_two-56_three-386_one
AMBER mailing list
Received on Sat Nov 06 2010 - 21:00:02 PDT
Custom Search