[gmx-users] PMF for Carbohydrate-Protein

[Justin A. Lemkul]

Sai Kumar Ramadugu wrote:
> Dear All,
> I'm running set of umbrella sampling simulations to get the PMF of a 
> disaccharide binding to a protein.
> I followed the tutorial provided by Justin and changed the values of 
> necessary parameters according to my system.
> The mdp file for umbrella sampling simulations is as follows:
> -------------------------------------------
> ; Pull code
> pull                    = umbrella
> pull_geometry   = distance      ; can't get PMF with direction
> pull_dim            = N N Y
> pull_start           = yes
> pull_ngroups    = 1
> pull_group0       = active_site
> pull_pbcatom0   = 575
> pull_group1       = new
> pull_pbcatom1   = 1500
> pull_init1             = 0
> pull_rate1           = 0.0
> pull_k1               = 750           ; kJ mol^-1 nm^-2
> pull_nstxout        = 1000          ; every 2 ps
> pull_nstfout         = 1000          ; every 2 ps
> ----------------------------------------------------
>  
> Active site in the pull_group0 being the residues in the binding site 
> and new in the pull_group1 being the disaccharide.
> Initially I did an SMD run where I separated the disaccharide from the 
> binding upto 1.5nm in distance.
>  
> When I started running the umbrella sampling simulations, I used the 
> conformations with approximately 0.1nm difference starting from an 
> initial separation of 0.3nm between the binding site residues and 
> disaccharide.
> My question is: when I run g_wham and get the profile, the pmf reaches a 
> plateau, but its not completely horizontal. It drops down a little. Is 
> it acceptable to have the pmf reach a plateau but not completely horizontal?

It looks like you just have a very rocky PMF, indicating that you probably 
haven't sampled sufficiently (i.e., the simulations are probably too short).

> In the histogram, the first two windows have good overlap and are narrow 
> distributions where as the later windows are spread out.
> I have tried to decrease the distance from 0.1 nm to 0.05 nm and still 
> the sampling around 0.5-0.75 of Xi region does not sample well.
> Is there anything that you could suggest?
>  

More windows and/or longer simulations.

-Justin

> I have uploaded the plots of profile and histograms.
>  
> PMF Plot link: http://img801.imageshack.us/i/pmf.png/
> Histogram link: http://img217.imageshack.us/i/histoy.png/
>  
> 
> Thanks for your time
>  
> Sai Ramadugu
> 

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

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

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