[gmx-users] Umbrella sampling windows

[Justin A. Lemkul]

Susana Tomasio wrote:
> Hi all,
> 
> I'm running umbrella sampling of an ion through a lipid bilayer with 
> gromacs 4.5.1.
> I used g_wham to create the histograms of the configurations within the 
> umbrella sampling windows (1 Angstrom interval).
> I did not get a sufficient overlap between the windows, so I was 
> wodering which is the better way of increasing the sampling:  to
>  include additional windows in the regions where there is no overlap or 
> to increase the force constant ?

If you increase the force constant, you will make the distributions narrower, 
and thus I would expect the overlap would be worse.  Insufficient simulation 
length could be an issue, too, but you haven't said how long your simulations are.

> If I increase the force constant can I continue the simulation with the 
> new constant or do I have to start again?

You'd have to start over again, I'd think, otherwise if you pass one .tpr file 
to g_wham per window, it will contain incorrect information that will mess up 
the calculations.

> I used a force contant of 3000 kJ mol^-1 nm^-2.
> 

That seems somewhat high, but there are no hard and fast rules about these 
things, I don't think.  You probably want either (1) a lower force constant or 
(2) more windows along your reaction coordinate.  Option (2) seems to be more 
efficient, since you don't have to re-do your simulations, you can just run some 
more.

-Justin

> Thank you in advance,
> 
> Susana
> 

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

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

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