[gmx-users] Issue with PMF

[Justin Lemkul]

On 4/5/20 7:13 AM, Quyen V. Vu wrote:
> Dear Justin,
> In your tutorial, you selected the initial configuration for each umbrella
> by the radial distance between chain A and B [ r = sqrt(x^2+y^2+z^2)]
> and in umbrella sampling simulation, your setting is:
> pull_coord1_dim = N N Y
>
> pull_coord1_start       = yes
>
> so as the draw I attached here you can see that conf1,conf2 and conf3
> are satisfied the distance criteria and you used the staring
> z-separation as reaction coordinates, this mean umbrella 1 and 2 will
> sample the same region while the region between zeta_1 and zeta_3 is
> not sampled well if |zeta3-zeta1| is large.
>
> https://drive.google.com/open?id=1nUKLmaBhjL29ZadgUcPk6DlG08ZVP_Nx
>
> Is it true?

Of course it is possible to draw infinitely many configurations that 
satisfy one dimension but not the total distance.

The tutorial uses COM distance primarily for generating reasonable 
starting configurations, and over the time scale used, the z-distance 
and COM distance should not differ much (I proved this to my self 10 
years ago when I first wrote the tutorial and did the underlying study). 
The restraint is applied along z only during the umbrella sampling and 
that is the reaction coordinate for which the free energy change is 
computed.

-Justin

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

Justin A. Lemkul, Ph.D.
Assistant Professor
Office: 301 Fralin Hall
Lab: 303 Engel Hall

Virginia Tech Department of Biochemistry
340 West Campus Dr.
Blacksburg, VA 24061

jalemkul at vt.edu | (540) 231-3129
http://www.thelemkullab.com

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