Minimum Distance CV
Suk Joon Na
As far as I know, minimum distance CV indicates the minimum distance
between two groups of atoms, specifically from VDW surface to surface
of atoms (i.e. not center of atoms).
I'm using this CV to get free energy profile of two plate structures
by Well-Tempered metadynamics.
From VMD, I can directly get the distance between these two plate
structures.
However, the distance in HILLS file shows very different values from
VMD.
For example, the upper plate travels up to 42.5 Aungstroms maximum
from the lower plate (it is allowed to move only in z direction) but
HILLS file shows it moves only 27 Aungstroms.
In other words, the minimum distance can vary from 0 to 42 A in real,
but PLUMED shows it moves only from 0 to 27A.
The PLUMED inpufile looks like this,
HILLS W_STRIDE 1000 HEIGHT 0.5
WELLTEMPERED SIMTEMP 300 BIASFACTOR 15
MINDIST LIST <g1> <g2> SIGMA 0.35 BETA 500
g1->
LOOP 1 36 1
g1<-
g2->
LOOP 37 72 1
g2<-
ENDMETA
So, can anyone explain what really happens in PLUMED?
Thank you in advance for your help.
Sincerely,
Joon
davide branduardi
I am not an expert in mindist variables but, if the atoms are put in a
sort of plate
and this plate is able to move up and down toward the other plate then
in the summation of the formula in mindist I expect that several terms
participate to the summation (there is no single term which is
remarkably dominant over the others in the sum of exponentials).
In this case
s=beta/(log sum_ij exp(beta/rij))
=beta/(log n_terms exp (beta/r))=beta/(log n_terms+ beta/r)
that would give a value which is remarkably lower than r
due to the effect of many competing terms.
Does this make any sense?
Ciao
Davide
> --
> You received this message because you are subscribed to the Google Groups "PLUMED users" group.
> To post to this group, send email to plumed...@googlegroups.com.
> To unsubscribe from this group, send email to plumed-users...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/plumed-users?hl=en.
>
>
--
**************************************************************
Davide Branduardi
Post Doc- Computational Chemistry
Drug Discovery and Development
ISTITUTO ITALIANO DI TECNOLOGIA
Via Morego 30, 16163 - Genova, Italy
Phone: +39 010 71781573 Room: 1/27
Fax: +39 010 7170187
Massimiliano Bonomi
Consider also that you are calculating distances with PBC,
so they might be lower than what VMD shows.
Try to add NOPBC on the line where you have the MINDIST keyword and
see what happens.
Max
Davide Provasi
if your 'plates' are always normal to the z direction,
you can use the distance variable projecting it on the Z direction
DISTANCE LIST <a> <b> DIR Z
this should also give the same order parameter
also notice that unless you specify not to with NOPBC, CV take into
account periodic boundary
conditions. is you box big enough so that this the minimum distance is
what you are interested in?
finally I don't think that your statement "specifically from VDW surface to
surface of atoms (i.e. not center of atoms)." is correct.
CV's are calculated from the atom coordinates
good luck
Davide
--
Davide Provasi
Dept. of Structural and Chemical Biology
Mount Sinai School of Medicine
Icahn Medical Institute Building
1425 Madison Avenue, Box 1677
New York, NY 10029-6574
Tel.:212-659-8618
Fax: 212-849-2456
Davide Provasi
obviously I didn't see massimiliano's reply before posting mine!
just consider that if you add NOPBC and keep a smaller box, you free-energy will
still include the effects from the interactions with the periodic
images, so be careful
when interpreting your results.
D
Giovanni Bussi
I think there is no reason to limit the choice to finite BETA. With
BETA=infinite the CV is still continuous and derivable, so it should
be suitable for being used in biased simulations (the derivative is
just discontinuous at a point, which is not a problem for MD; same as
truncated LJ potentials). In my opinion, BETA=infinity could even be
the default! But I admit that I never used it, so I am definitely not
an expert on that.
Giovanni
On Sat, Jun 11, 2011 at 6:59 PM, Davide Provasi
Suk Joon Na
I tested the model with NOPBC, however, the distance is still quite
different from VMD.
So, I modified BETA from 500 to 5000, then the distance shows
approximately correct values! ( -+ 1~2 aungstroms).
However, somewhat the simulation got errors during simulation, and I
hope to believe these error are due to other problems, not PLUMED.
(I'm using LAMMPS, and it gives 'nan' errors after hundreds of ps.
It's funny when I apply the BETA 500, it seems working well)
I will post the results if I figure out what the problem is due to.
Again, I appreciate your great helps.
Sincerely,
Joon
On Jun 13, 6:13 am, Giovanni Bussi <giovanni.bu...@gmail.com> wrote:
> Hi guys, I add another comment:
>
> I think there is no reason to limit the choice to finite BETA. With
> BETA=infinite the CV is still continuous and derivable, so it should
> be suitable for being used in biased simulations (the derivative is
> just discontinuous at a point, which is not a problem for MD; same as
> truncated LJ potentials). In my opinion, BETA=infinity could even be
> the default! But I admit that I never used it, so I am definitely not
> an expert on that.
>
> Giovanni
>
> On Sat, Jun 11, 2011 at 6:59 PM, Davide Provasi
>
> >>> --
> >>> **************************************************************
> >>> Davide Branduardi
> >>> Post Doc- Computational Chemistry
> >>> Drug Discovery and Development
> >>> ISTITUTO ITALIANO DI TECNOLOGIA
> >>> Via Morego 30, 16163 - Genova, Italy
> >>> Phone: +39 010 71781573 Room: 1/27
> >>> Fax: +39 010 7170187
>
> >>> e-mail:
> >>> davide.brandua...@gmail.com
> >>> --
> >>> You received this message because you are subscribed to the Google Groups "PLUMED users" group.
> >>> To post to this group, send email to plumed...@googlegroups.com.
> >>> To unsubscribe from this group, send email to plumed-users...@googlegroups.com.
Davide Branduardi
I would like to add a small bit.
If you use a very high beta then the force will be applied only on the atom which is actually at lowest distance.
In case you have another atom that becomes the one at lowest distance this immediately experience a huge
force on it which probably may be too traumatic for your system (namely: your integrator does not follow the abrupt change
in forces). I would use the BETA 500 value if this work that makes your exponential smoother. The fact that this is not "exactely" the min dist is somehow
not necessary. You just need something that is proportional to it and I have the impression that, even with beta 500 you
may obtain a good order parameter for your transition.
Ciao
Davide
*********************************************
Davide Branduardi
Postdoc
Theoretical Molecular Biophysics Group
Max Planck Institute for Biophysics
Max-von-Laue-Straße 3
60438 Frankfurt am Main
Phone: +49 069 63031505
Fax: +49 069 63031502
davide.b...@biophys.mpg.de
davide.b...@gmail.com
*********************************************