[gmx-users] PMF profile from Umbrella - Plateau

[Justin A. Lemkul]

On 5/16/12 10:25 AM, Steven Neumann wrote:
> Dear Justin,
>
> I pulled my ligad away of 6nm from the protein and obtained beautiful and smooth
> curve of PMF using 28 windows. Starting from zero kcal/mol corresponding to app
> 0.3 nm then minima and curve increase till my plateau which starts from app 1 nm
> - then I have nearly straight line till 6.3 nm.
> My question: Can I decrease pulling distance to save time for the future
> simualtions with this system in terms of number of water molecules? Shall I just
> pull it away of 2-3 nm and I will obtain the same deltaG?
>

Ideally, one would separate the molecules such that they are no longer 
interacting.  Obviously with methods like PME this is impossible, so what you're 
after is sufficient separation such that there is negligible interaction between 
the protein and ligand.  The required distance depends on the nature and size of 
the ligand, the types of interactions it experiences, and the cutoffs used. 
Keep in mind that even if a molecule is beyond the longest nonbonded cutoff, 
there are still effects like water ordering that can persist up to about 1.0 nm, 
so I would certainly make sure that all the atoms of the ligand and protein are 
separated by a minimum of twice the longest cutoff or so to account for this 
effect.  Note that this would imply the use of g_mindist, not g_dist, as the 
distance I'm talking about is not the same as the COM distance.

That's just my own rule of thumb for an initial check; you'll have to analyze 
your own system to decide what's happening and what you should do.

-Justin

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

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

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