# [gmx-users] Distance of pull group is larger than half the box size

Justin A. Lemkul jalemkul at vt.edu
Sun Jun 17 14:40:12 CEST 2012

```
On 6/17/12 7:47 AM, Eudes Fileti wrote:
> Hello guys, I have tried to calculate the PMF for a dimer in a
> rectangular box of dimensions 3 x 3 x 7. However, the simulation
> stops when the COM separation is above 1.49 (ie half
> of the smallest dimension of the box. Since I am pulling along
> the Z axis at a distance that would be up to 3.5 nm.
>
> I've tried both pull_geometry, distance and direction, to try to get
> around this, but I have not been successful.
>
> Do I have to use the dimensions of the box as twice the COM
> distance (in my case: 5 x 5 x 5; to give me r_max = 2.5nm)?
>
> Some of you could give me a hand?
>
>
> DIRECTION
> pull                     = umbrella
> pull_geometry            = direction
> pull_dim                 = Y Y Y
> pull_group0              = mol_0
> pull_group1              = mol_1
> pull_rate1               = 0.01
> pull_vec1                = 0.0 0.0 1.0
> pull_k1                  = 3000
>
> DISTANCE
> pull                     = umbrella
> pull_geometry            = distance
> pull_dim                 = Y Y Y
> pull_group0              = mol_0
> pull_group1              = mol_1
> pull_rate1               = 0.01
> pull_vec1                = 0.0 0.0 0.0
> pull_k1                  = 3000

If you only want to pull along the z-axis using the "distance" geometry, then
you should not set "Y Y Y" for the pull_dim here; that will cause pulling along
all 3 axes, hence why you're having problems in x and y.  For z-only, set "N N Y."

It is not clear to me why the settings for the "direction" pulling do not work.
In that case, the settings for pull_dim are ignored and pull_vec1 is used.
There is a different setting that should be more appropriate.  Read in the
manual about the "direction_periodic" geometry.  It overcomes these issues.
Otherwise, your box needs to be at least twice the COM distance at all times.

-Justin

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

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

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

```