[gmx-users] Re: reflecting sphere

[Pim Schravendijk]

> From: Wang Zhun <wangzhun76 at gmail.com>

> Hi, Dr Spoel,
> I wonder if a reflecting sphere can be created by Gromacs. In  PNAS
> 2004,101(35):12916-12921, Charmm was used to created a reflecting
> sphere to restrict the interaction simulation between 2 peptide
> chains. Can gromacs do the same job?
> By the way, several posts sent from my email box turned to be
> irregular characters in the mail list. I don't know how this happened,
> because I could receive and read  these my own posts in English in my
> email box.
> Thank you for your help in advance.

There were some answers to your question before, did you get any further 
with that?

http://www.gromacs.org/search/index.php?query=reflecting+sphere&idxname=mailing_lists

Their setup of the wall is described in the suppl. information:
 
http://www.pnas.org/cgi/content/full/0402634101/DC1/1 (registration required)

As you can see they just defined a harmonic potential with a force
constant of 1kcal/(mol*A^2) applying on all particles that move outside
the radius. You could do the same by calculating the resulting force and
adding this to the force array. Put it in gmx/src/kernel/md.c, for example
just before the AFM and Umbrella type pulling is called. It needs the x
coordinates from the "state->x"  array, and the calculated force should be
added per atom to the "f" array.  If you define the center of the sphere
in 0,0,0 and since you don't use PBC, you probably don't need to know the
box coordinates. The only tricky bit is that you should give the direction
of the force along the local normal of the spherical surface.

Probably it will only be a few lines that you would have to add! 

Greetings, Pim