Re: Spinangle with refgroup : Warning: discontinuous rotation!

From: Giacomo Fiorin (giacomo.fiorin_at_gmail.com)
Date: Mon Apr 07 2014 - 19:54:30 CDT

Hello Nicolas, for this to work you need to be sure that both the group
defining the spinangle and the reference group that defines the rotating
frame do not change internal structure too much.

By "too much" I mean that the RMSD of either group doesn't exceed the
radius of gyration (as a rule of thumb).

These dramatic changes of structure may occur fairly easily in a system in
vacuo such as deca-alanine. If your ultimate goal is to use a reference
group to define a rotating frame but not to apply a restraint to it, my
suggestion would be to make the reference group and the "restrained" group
be completely different molecules.

Giacomo

On Mon, Apr 7, 2014 at 12:32 PM, nicolas martin
<nicolasmartin973_at_gmail.com>wrote:

> Dear NAMD users,
>
> As part of my project, I planned to compute the spinangle of the upper
> part of the protein and set the lower part as a reference. To develop my
> method i chose to work with deca-alanine working with one extremity,
> computing the spinangle of the last part of ALA10 around an axis, defined
> as the axis going through one bound nearby. This spinangle is then
> restrained to obtain a rotation of the mobile part of my selection. Exactly
> as one would do computing a dihedral angle and restraining it to explore
> [-180;180[.
>
> I tried to be as clear as possible to explain what I aim to do, but I am
> sure that is no trivial thing to get. Anyway my point by seeking for help
> is that I cannot success in running long simulations of this system. For
> the short ones, I got some "Warning: discontinuous rotation!" but the
> simulation can end properly. When I try to elongate the simulation, the
> velocities of random atoms happens to be too high so does the temperature
> of the system and the calculation crash before term.
>
> I figured out by deduction that this "discontinuous rotation" was due to a
> bestfit step (which could occur in both spinangle and refgroup) but since I
> do not really understand why I might be mistaken. Maybe this is the way I
> want to restrain and compute this spinangle which is totally wrong. You
> will find below the definition of my colvar.
>
> Hoping you can help me on this one, bests,
>
> Nicolas Martin
> Laboratoire d'Ingenierie des Fonctions Moleculaires
> Universite de Strasbourg
>
> colvar {
> name spinangle
> spinangle {
> atoms {
> atomNumbers 92 93 94 95 96 99 100 101 102 103 104 # measure group
> centerReference on
> rotateReference on
> refPositionsGroup { # reference group
> atomNumbersRange 1-91
> }
> refPositionsFile input/ref.pdb
> refPositionsCol O
> refPositionsColValue 1.00
> }
> refPositionsFile input/mesure.pdb
> refPositionsCol O
> refPositionsColValue 1.00
> axis (-0.476,0.644,1.189) # rotation axis defined along a bond
> }
> }
> harmonic {
> name harmonic
> colvars spinangle
> forceConstant 0.0060923 # this should be equal to 10kcal/mol/radian
> centers -180
> targetCenters 180
> targetNumSteps 50000
> targetNumStages 37
> }
>
> colvar {
> name rmsd
> rmsd {
> atoms { atomNumbersRange 1-91 # reference part is restrained by RMSD
> }
> refPositionsFile input/ref.pdb
> }
> }
> harmonic {
> colvars rmsd
> forceConstant 3000
> centers 0
> }
>
>

This archive was generated by hypermail 2.1.6 : Thu Dec 31 2015 - 23:20:40 CST