[gmx-users] umberella sampling

[Justin A. Lemkul]

mohsen ramezanpour wrote:
> Dear All and Specially Dear justin
> 
> After reading  umberella sampling tutorial I have afew question.
> Please let me know their answers.
> 
> 1-How did you realized(What was your criteria) the range of 0.5 -5 nm is 
> enough? it might be different if you chose 0.2-5 nm.Am I right?
> 

A PMF is measured over a reaction coordinate.  If you change the reaction 
coordinate, you're measuring something different.  For the system in the 
tutorial, 0.2 nm of COM spacing is not possible.  The beta strands cannot be 
force any closer together.

> 2-The histogram showes a non-overlaping for 1-3 windows,Do I need to run 
> all  23 simulations again  with more windows (and small spcing)this time?
>    what is the solution to gain an exact estimation of PMF?
>   

No.  If windows do not overlap, then you can either:

1. Extend the simulations in the non-overlapping windows.
2. Add more windows between those that do not overlap.
3. Use a different force constant to allow more mobility within a window.

> 3-Can I select windows in different spacing?in the other words,can I 
> choose small spacing for 1-3 windows and 0.2 nm for others as you chose?
> 

Please refer to the paper cited in the tutorial.  We used uneven spacing to 
actually collect the data, and the reasons for it are discussed.

-Justin

> Thanks in advance for your reply
> 
>  
> 

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

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

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