Re: Colvars periodic boundary treatment question

From: Kasra Fattah (kasra.fattah_at_gmail.com)
Date: Tue Dec 03 2013 - 15:57:04 CST

Jerome,
Using "distnaceXY" and assuming that I want the molecule to remain within
e.g. 5 from the ref point, how should I use the "upperWall" and "lowerWall"
values? Is that just: "upperwall 5"? As I understand "distanceXY" colvar is
a norm and is just positive so I can only use positive values for
"upperWall" and "lowerWall" right?

Cheers,
Kasra.

On Tue, Dec 3, 2013 at 12:39 PM, Jrme Hnin <jerome.henin_at_ibpc.fr> wrote:

> The natural complement to distanceZ for such a case is distanceXY. You can
> restrain that.
>
> Jerome
>
> ----- Mail original -----
>
> > Thanks for your response Jerome.
> > I also tried using "period" and "wrapAround" options with "period 50" and
> > "wrapAround 0.0" (basically the default value of wrapAround) in distanceZ
> > and I used the unwrapped coordinates for the walls positions, i.e.
> > "lowerWall 20" and "upperWall 30". It seems that it's working too, is
> that a
> > legit treatment of the problem or not? (btw, I'm using a dummy atom as a
> > reference group at position (0.0, 0.0, 0.0) )
>
> > One more question:
> > I'm actually doing a 1D ABF calculation and my reaction coordinate is Z
> > direction of the cartesian system. But in order to limit the simulation
> in x
> > and y directions I'm defining two other reaction coordinates, x and y.
> The
> > sole application of these two reaction coordinates are for using their
> > "lowerWall" and "upperWall" options to enforce the molecule remains in
> the
> > desired window in x and y directions but I'm not using them in free
> energy
> > calculation part of abf, i.e. just using "colvars z" in the abf part of
> the
> > config file. I was wondering if this is a right way of confining the
> > molecule in the directions that I'm not doing free energy calculations?
>
> > Cheers,
> > Kasra.
>
> > On Sat, Nov 30, 2013 at 7:04 PM, Jrme Hnin < jerome.henin_at_ibpc.fr >
> wrote:
>
> > > Hi Kasra,
> >
>
> > > The best way for this is to use distanceZ with a reference group close
> to
> > > the
> > > region of interest (if you want an absolute position along the axis, it
> > > would be a dummy atom group, and you can choose its position
> arbitrarily).
> >
> > > In the case you describe in your message, you could define a distanceZ
> > > component with the reference group as a dummy atom at (0, 25, 0), in
> which
> > > case the lowerWall would be -5 and upperWall would be 5. You don't
> need to
> > > worry about PBC, as they will be enforced automatically, because the
> > > distance between your molecule and the dummy atom will be computed
> > > following
> > > the minimum image convention.
> >
>
> > > That's why the distanceZ coordinate is not explicitly treated as
> periodic
> > > by
> > > default: the minimum image convention for distances usually does the
> job
> > > well.
> >
>
> > > Cheers,
> >
> > > Jerome
> >
>
> > > > Hi All,
> > >
> >
> > > > I've got a question regarding the treatment of a reaction coordinate
> that
> > > > crosses a periodic boundary.
> > >
> >
> > > > If I want to confine the center of mass of a molecule between two
> points
> > > > along, let's say, y axis which crosses the periodic boundary in that
> > > > direction how should I define lowerWall and upperWall? for example
> if the
> > > > box size in y direction is from -25 to 25 and initially the com of
> the
> > > > molecule is located at y=25 but I want it to be confined in
> y-direction
> > > > +/-
> > > > 5 from its initial position. Based on unwraped values it should be
> > > > between
> > > > 20 and 30, i.e. "lowerWall 20" and "upperWall 30" but I'm not sure
> if I
> > > > have
> > > > to use unwrapped coordinates or wrapped ones?!
> > >
> >
>
> > > > I Would appreciate it if you could help me to know how I should set
> these
> > > > values...There are also two options in the manual "period" and
> > > > "wrapAround"
> > > > should I also set them (i.e. "period 10" and "wrapAround 25")? or
> they
> > > > are
> > > > not related to what I need?
> > >
> >
>
> > > > Cheers,
> > >
> >
> > > > Kasra.
> > >
> >
>

This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:24:03 CST