[gmx-users] large error bars in PMF

[Justin A. Lemkul]

Justin A. Lemkul wrote:
> 
> 
> Rebeca García Fandiño wrote:
>> Hi,
>> thanks a lot for your quick answer.
>> What I am trying to pull are two small peptides one from another (r_1 
>> and r_2).
>> I did not understand very well your last suggestion: "...if you want 
>> reasonable error bars you will not lots of well-converged data".
> 
> Oops, that should have read "you will *need* lots of well-converged data."
> 
>> Do you mean I will need also more windows besides extending the 
>> simulations?
> 
> I doubt you need more windows.  Likely you just need more time in each.
> 
>> I think the problem could be also that the peptides I am using rotate 
>> in the box and they do not remain flat one respect to the other. They 
>> gyrate freely and some parts of their structure interact along the 
>> pulling...
> 
> Interactions are part of the dissociation process and are not 
> problematic per se.  But if you're trying to obtain only a 
> one-dimensional PMF then your rotation could be a problem.  Is there 
> some reason you need a one-dimensional PMF and not a three-dimensional 
> PMF?  What are you trying to achieve?
> 

I should also add that if you're getting spurious interactions across periodic 
boundaries, this is a problem.  In that case, you need a different type of box, 
like a dodecahedron for spherical symmetry.

-Justin

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

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

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