[gmx-users] Distance restraints does not work?

[Justin Lemkul]

On Thu, Dec 12, 2013 at 8:22 AM, Emma Ahlstrand <emma.ahlstrand at lnu.se>wrote:

> Hi,
> I am trying to do a simulation with distance restraints for the
> Zn-His(NE2)atoms in a peptide, but the distances are not kept to what I
> want them to be. Why?
>
> I have added this in my topology.top:
> ; Include Position restraint file
> #ifdef POSRES
> #include "posre.itp"
> #endif
>
> #ifdef ZN
> [ distance_restraints ]
> 355    333      1       1       1       0.1985  0.2231  0.4     1
> 355    316      1       2       1       0.1985  0.2231  0.4     1
> 355    156      1       3       1       0.1985  0.2231  0.4     1
> 355    139      1       4       1       0.1985  0.2231  0.4     1
> #endif
>
> And I use the following setting in pr.mdp:
> define              =  -DZN
> ; NMR refinement stuff
> disre           = Simple        ; Distance restraints type: No, Simple or
> Ensemble
> disre-weighting = Equal         ; Force weighting of pairs in one dist
> rest: Equal or Conservative
> disre-mixed     = no            ; Use sqrt of the time averaged times the
> instantaneous violation
> disre-fc        = 1000
> disre-tau       = 0
> nstdisreout     = 100           ; Output frequency for pair distances to
> energy file
>
> When I look at the structure, two of the restrained His(atom 139 and 333)
> has left the Zn(atom 355). Some values from the distance analysis are
> printed below:
> g_dist -f traj.trr -s b4pr.tpr -n index.ndx -o dist139.xvg
> less dist139.xvg
> @    xaxis  label "Time (ps)"
> @    yaxis  label "Distance (nm)"
> @ s0 legend "|d|"
>    0.0000000    0.2119962
>   24.0000000    0.2470244
>   50.0000000    0.4642269
>  100.0000000    0.4505744
>
> less dist333.xvg
> @    xaxis  label "Time (ps)"
> @    yaxis  label "Distance (nm)"
> @ s0 legend "|d|"
>     0.0000000    0.2172887
>     24.0000000    0.2171661
>     50.0000000    0.4632832
>  100.0000000    0.3777701
>
> Why does the distance increase? Why is it not held in between roughly
>  0.1985  0.2231?
>
>
A restraint is a penalty.  It does not absolutely fix certain quantities,
so the distances can deviate.  This is especially true when you have
multiple restraints that must be simultaneously satisfied - they compete
with one another.  Note the functional forms in the manual; your settings
indicate that no penalty is applied between 0.1985 and 0.2231 nm, and only
after 0.4 nm is a large penalty imposed.  So there's a fair amount of
wiggle room there for each restraint.  You can adjust the coefficients to
produce a better result, or simply introduce constraints for these "bonds,"
in which case the originally geometry will be preserved exactly.

-Justin

-- 

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

Justin A. Lemkul, Ph.D.
Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441


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