Re: [plumed-users] Re: failure of lower_walls and upper_walls

[Giovanni Bussi]

Hi,

I think the problem is that torsion is periodic. In the range between pi and 2pi you will have both walls acting.

Can you try applying a single lower wall at zero to the sin of the angle?

Giovanni

 

On Wed, Aug 9, 2017 at 9:28 AM, <qingh...@gmail.com> wrote:

Just an additional information, the wall applied on distance cv works for me, but not for the torsions. Thanks!


Qinghua

On Wednesday, August 2, 2017 at 11:40:00 AM UTC+2, qingh...@gmail.com wrote:

Hello PLUMED users,

I have tried to apply the lower and upper walls in a metadynamics simulations using a torsional angle as CV, but I found that the walls were failed.

Here is my input file for plumed:

    UNITS LENGTH=A TIME=ps ENERGY=kcal/mol

    chain: GROUP ATOMS=1-3470
    WHOLEMOLECULES ENTITY0=chain

  dihe10C: TORSION ATOMS=168,160,162,165

  uwall: UPPER_WALLS ARG=dihe10C AT=pi  KAPPA=150.0 EXP=2 EPS=1 OFFSET=0
  lwall: LOWER_WALLS ARG=dihe10C AT=0.0 KAPPA=150.0 EXP=2 EPS=1 OFFSET=0

 metad: METAD ARG=dihe10C PACE=500 HEIGHT=0.3 SIGMA=0.3 BIASFACTOR=16.0 TEMP=300.0 FILE=HILLS GRID_MIN=0 GRID_MAX=pi GRID_BIN=100
  PRINT STRIDE=50 ARG=dihe10C,metad.bias,uwall.bias,lwall.bias FILE=colar.dat


I only want to investigate the profiles between 0 and pi for the torsion, and the starting CV fell in the range. But after some steps, the CV went beyond the range.

The CVs stay at around -1.6 radian, though the bias potential grows rapidly and the wall potentials stay high. This is unreasonable, and I could not figure out the problem.

Could some one give me some suggestions? Thanks a lot!


All the best,
Qinghua

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users+unsubscribe@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at https://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/4e3032a3-216a-4ff7-85c2-db1320a2bd62%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[qingh...@gmail.com]

Dear Prof. Bussi,

Thanks a lot for your explanation and suggestion.
I was suspecting that the reason might be the periodicity of torsion.

Finally, I worked it out by setting walls to sin of the torsion.


PS: sorry for deleting the first post. I thought that it might be a very simple question to other people.

All the best,
Qinghua

To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at https://groups.google.com/group/plumed-users.

[sagarb...@gmail.com]

Hey

I am also facing similar kind of situation I am using distance in z direction as CV and I am applying lower wall at -6.0 and upper wall at +6.0.
Is it fine ???
My plumed.dat file looks like this
first:COM ATOMS=1-5402
second:COM ATOMS=20751-20765
comdist: DISTANCE ATOMS=first,last
COMPONENTS

lwall: LOWER_WALLS ARG=comdist.z AT=-6.0 KAPPA=1000 EXP=2.0 EPS=0.5 OFFSET=0.0
uwall:UPPER_WALLS ARG=consist.z AT=6.0 KAPPA=1000 EXP=2.0 EPS=0.5 OFFSET=0.0

meta_grid: METAD ARG=comdist.z PACE=500 HEIGHT=0.2 SIGMA=0.2 FILE=HILLS_GRID GRID_MIN=-6.0 GRID_MAX=6.0 GRID_SPACING=0.04

PRINT STRIDE=10 ARG=comdist.z,metad.bias FILE=COLVAR_GRID

ENDPLUMED

I got an error that cv is looking for a value outside the grid point but it shouldn't be becoz of walls
Please help

Sagar

[qingh...@gmail.com]

Hello,

As you set the GRID_MIN and GRID_MAX exactly the same as the lower and upper walls, respectively.

There might be still some points going out of the walls, which will lead to grid failure.

You can set the GRID range a little larger than the wall range, it can solve the problem.

All the best,

Qinghua

[sagarb...@gmail.com]

Thanks for replying
And I will apply and check as u suggested
Why CV is crossing the value beyond wall
I can't understand that.
Please throw some light on it

One more thing , when periodicity affects the position of wall or anything??

[qingh...@gmail.com]

Hello,

You can check your CVs during the simulation. Even there is a high wall potential,

there is still a possibility to cross the wall, though it will be pull back by the potential.

I am not sure the periodicity will affects your simulations, it depends on your box size

and the distance range you want to sample along component Z.

You can have a check on the keyword DISTANCE with COMPONENTS.

All the best,

Qinghua

[sagarb...@gmail.com]

Ok thank you so much, I will do that too
In second case, I am using 2 CV's
Plumed.dat file looks like this

first:COM ATOMS=1-5402
second:COM ATOMS=20751-20765

g1_lipid: COM ATOMS=1-1274,2549-3244,3999-4700
g2_lipid:COM ATOMS=1275-2548,3245-3998,4701-5402

comdist: DISTANCE ATOMS=first,last
COMPONENTS

angle: ANGLE ATOMS=g1_lipid,g2_lipid,20757,20760

lwall: LOWER_WALLS ARG=comdist.z AT=-6.0 KAPPA=1000 EXP=2.0 EPS=0.5 OFFSET=0.0
uwall:UPPER_WALLS ARG=consist.z AT=6.0 KAPPA=1000 EXP=2.0 EPS=0.5 OFFSET=0.0

meta_grid: METAD ARG=comdist.z,angle PACE=500 HEIGHT=0.2 SIGMA=0.2,0.35 FILE=HILLS_GRID GRID_MIN=-6.0,0 GRID_MAX=6.0,3.1415 GRID_SPACING=0.04,0.07

PRINT STRIDE=10 ARG=comdist.z,angle,metad.bias FILE=COLVAR_GRID

ENDPLUMED

Do I need to consider periodicity and walls for the angle

[qingh...@gmail.com]

Hello,

On Monday, June 4, 2018 at 12:59:54 PM UTC+2, sagarb...@gmail.com wrote:

comdist: DISTANCE ATOMS=first,last
COMPONENTS

angle: ANGLE ATOMS=g1_lipid,g2_lipid,20757,20760

 

It seems that you set a torsion here, but not a angle.

 

lwall: LOWER_WALLS ARG=comdist.z AT=-6.0 KAPPA=1000 EXP=2.0 EPS=0.5 OFFSET=0.0

PRINT STRIDE=10 ARG=comdist.z,angle,metad.bias FILE=COLVAR_GRID

ENDPLUMED

Do I need to consider periodicity and walls for the angle

I guess that you don't need to consider the PBC for the angle (torsion) once you set the atoms involved in the CVs as a whole group.

Then PLUMED will take care of it.

All the best,

Qinghua

[sagarb...@gmail.com]

Thank you so much for your help
Actually, in this case I am using angle calculated by 4 atoms
First 2 atoms will represent a vector and last 2 another
Then plumed calculated the angle between them

Best
Sagar