[gmx-users] refcoord-scaling

[Mark Abraham]

On Apr 8, 2014 12:14 AM, "Chetan Mahajan" <chetanvm10 at gmail.com> wrote:
>
> Hi,
>
> Thanks a lot. It is clearer, but I have questions, so this email:
>
> Why is option "none" not desirable, if we want to restrain positions of
> atoms of crystal to specific points? Isn't the strain generated due to
> atoms scaling but not reference coordinates helping?

Only if the model can absorb that stress at the given time step, which
might not be true with heavily coupled constraints.

> Why is none option
> better than "com"? Is the possible movement of crystal due to "com" option
> desirable?

I already offered a theory about com not working in your case.

> If a particular option exists, then it does so for a particular purpose.
In
> which case, "none" or "all" be a valid option?

These are all unphysical hacks, so "it was easy to implement and might be
useful enough for something" is a valid purpose when building a general
purpose tool. Probably protein use cases were considered, and not crystal
structures. Do something that works and move on to real equilibration! :-)

Mark

> Thanks
> Chetan
>
>
> On Mon, Mar 24, 2014 at 10:40 PM, Christopher Neale <
> chris.neale at alum.utoronto.ca> wrote:
>
> > Say you have a box with an x-side length of 3 nm and two atoms with
> > x-dimension position restraints to: (a) x=1 nm, and (b) x=2 nm
> >
> > Then let the box shrink to 99% of its previous size due to pressure
> > coupling. The following is my understanding of  the locations to which
the
> > atoms will be restrained:
> >
> > refcoord_scaling =
> >
> > none:
> > (a) x=1 nm, and (b) x=2 nm  -- note how the restraint positions are not
> > scaled (even though the positions of the atoms are scaled due to
pressure
> > coupling)
> >
> > all:
> > (a) x=0.99 nm, and (b) x=1.98 nm   -- note how the atoms are now
> > restrained to be closer together
> >
> > com:
> > (a) x=0.985 nm, and (b) x=1.985 nm  -- note how the distance between the
> > atoms is maintained
> >
> >
> > So you see how "none" leads to problems because the restraint reference
> > positions are not scaled at all with the atomic positions and how "all"
> > leads to problems because, if you restrained all Ca atoms, then you
would
> > get compression/expansion of the size of your protein as the system
> > fluctuated in constant pressure simulations.
> >
> > Chris.
> > ________________________________________
> > From: gromacs.org_gmx-users-bounces at maillist.sys.kth.se <
> > gromacs.org_gmx-users-bounces at maillist.sys.kth.se> on behalf of Chetan
> > Mahajan <chetanvm10 at gmail.com>
> > Sent: 24 March 2014 23:25
> > To: gmx-users at gromacs.org
> > Subject: Re: [gmx-users] refcoord-scaling
> >
> > Thanks, Justin. I am trying to understand what is meant by 'com' option.
> > Following two sentences in manual seem conflicting with each other:
> >
> > "*Scale* the center of mass of the reference coordinates with the
scaling
> > matrix of the pressure coupling. The *vectors of each reference
> > coordinate*to the center of mass are *not
> > scaled*."
> >
> > Doesn't scaling center of mass of reference coordinates change something
> > else ? reference coordinates?
> >
> > Thanks
> >
> > Chetan
> >
> >
> > On Mon, Mar 24, 2014 at 8:22 PM, Justin Lemkul <jalemkul at vt.edu> wrote:
> >
> > >
> > >
> > > On 3/24/14, 9:05 PM, Chetan Mahajan wrote:
> > >
> > >> Hi Justin,
> > >>
> > >> My system is TiO2 crystal (2160 atoms position restrained) solvated
by
> > >> 3656
> > >> water molecules, 1 formate anion and 1 sodium ion.  Is it true that
when
> > >> 'all' option is used, positions of the atoms (meant to be restrained)
> > >> always change?
> > >>
> > >>
> > > Consider what position restraints are doing.  They never guarantee
that
> > > atoms won't move; they just apply a biasing potential to disfavor
> > movement.
> > >  So yes, your atoms will probably move.  The smaller the system, the
> > larger
> > > the effect will likely be.  Consider also what reference coordinates
are
> > > doing.  They define how strong the potential is (distance between
current
> > > coordinates and reference).  If you're scaling all atoms individually
> > > according to the pressure coupling, the reference moves and therefore
the
> > > atoms can also move.  For larger systems like proteins,
> > "refcoord_scaling =
> > > com" is normal and stable, and "refcoord_scaling = all" is disfavored.
> > >  Honestly, I don't know what to expect for a very small system like
> > yours,
> > > but it is likely to be affected to a greater extent.
> > >
> > > If the "all" option gives you a stable simulation that allow for
proper
> > > equilibration, then I doubt there is any real problem.
> > >
> > >
> > > -Justin
> > >
> > > --
> > > ==================================================
> > >
> > > Justin A. Lemkul, Ph.D.
> > > Ruth L. Kirschstein NRSA 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
> > > http://mackerell.umaryland.edu/~jalemkul
> > >
> > > ==================================================
> > > --
> > > Gromacs Users mailing list
> > >
> > > * Please search the archive at http://www.gromacs.org/
> > > Support/Mailing_Lists/GMX-Users_List before posting!
> > >
> > > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> > >
> > > * For (un)subscribe requests visit
> > > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> > > send a mail to gmx-users-request at gromacs.org.
> > >
> > --
> > Gromacs Users mailing list
> >
> > * Please search the archive at
> > http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> > posting!
> >
> > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> >
> > * For (un)subscribe requests visit
> > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> > send a mail to gmx-users-request at gromacs.org.
> > --
> > Gromacs Users mailing list
> >
> > * Please search the archive at
> > http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> > posting!
> >
> > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> >
> > * For (un)subscribe requests visit
> > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> > send a mail to gmx-users-request at gromacs.org.
> >
> --
> Gromacs Users mailing list
>
> * Please search the archive at
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
send a mail to gmx-users-request at gromacs.org.