[gmx-users] questions about umbrella sampling tutorial

[Justin Lemkul]

On 4/5/13 3:29 AM, Albert wrote:
> Hello:
>
>   I am trying to do umbrella sampling for a protein/ligand water system these
> days following Justin's nice tutorial. And I've got some questions:
>
> (1) in the tutorial, the small peptide was pulling along Z direction. I am just
> wondering, will the small peptide be pulled in both +Z and -Z direction?
>
> In my case I've got a ligand in the deep pocket, it is possible that the ligand
> can escape the binding pocket from two opposite site, let's say -Z and +Z which
> I hope the umbrella sampling can give me some evidences to prove this....  Could
> the umbrella sampling answer this question? Or we should pull the ligand along
> X,Y and Z three direction so that it is more convinced?
>

The peptide is pulled/restrained in whatever direction you tell it.  The 
tutorial is just a simple example of the most trivial case, a one-dimensional 
pull along an axis.  You can do far more complex things using appropriate settings.

> (2) it is said that when doing free energy calculation in gromacs, we'db better
> use double precision Gromacs. So probably we have to use CPU instead of GPU for
> this?
>

I don't think the pull code works on GPU (someone correct me if I'm wrong), but 
in any case, I think the general reference to "free energy calculations" usually 
just refers to alchemical transformations.  In most cases with umbrella 
sampling, your error bars will be orders of magnitude larger than any increase 
in precision.

-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

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