Hey

I may need a bit more explanation to understand last two points you make..

*- csg_fmatch --trj traj.trr --force-trj traj_rerun_cg.trr (fill give*
*error force)*
*
*
*::-* will the output of this be .trr file with error forces or CG potential
which reproduces error forces. Also I don't see the --force-trj option in
help text/manual for csg_fmatch.

*You could test whether you did it correct be using force-matching*
*potentials as input for rerun -> error force should vanish*
*
*
This point is not clear.

thanks
sikandar

On Thu, Mar 31, 2011 at 11:19 AM, Victor Ruehle <[email protected]> wrote:

> Hey
>
> unless you change your basis set for cg force field (cut-off, add
> interactions, etc) or add some bias to fit procedure (extension to
> code needed), you will probably end up with identical potentials as
> doing only force matching.
>
> However, technically  it can be done the following:
> - csg_map --in traj.trr --out traj_cg.trr --cg mapping.xml (same for
> conf.gro)
> - prepare coarse-grained topolology with already known potentials
> - mdrun -rerun traj_cg.trr   (trajectory will contain knowledge based
> forces)
> - csg_fmatch --trj traj.trr --force-trj traj_rerun_cg.trr (fill give
> error force)
>
> You could test whether you did it correct be using force-matching
> potentials as input for rerun -> error force should vanish
>
> Victor
>
> 2011/3/31 Sikandar Mashayak <[email protected]>:
> > Hi
> > Another similar situation I am thinking of is like follow ::
> > 1. I have full atomistic trajectory and forces as a target system.
> > 2. I want to Coarse-Grain and determine CG-CG potential. I also have some
> > knowledge what this CG-CG looks like and and insist on using this
> > information. To do that I plan to do following.
> > 3. Use this knowledge based ( so to call) CG-CG to perform MD-Rerun with
> > full atomistic trajectories and get CG forces.
> > 4.  Now compute the error in forces on CG site for full MD and CG
> MD-rerun.
> > 5. Use this error force as target force to do Force-Matching and get an
> > estimate for CG-CG potential which reproduces this error.
> > 6. Finally add this error CG potential to original (knowledge based)
> > potential to get more accurate CG-CG potential.
> >
> > So the question I have is, how do I generate trajectories for just error
> > forces. I mean , is there any way I use two gromacs trajectories and get
> new
> > trajectory with error force?
> > thanks
> > sikandar
> >
> > On Mon, Mar 28, 2011 at 5:35 PM, Sikandar Mashayak <[email protected]
> >
> > wrote:
> >>
> >> thanks victor, I will see how subtracting known parts and setting
> interval
> >> to the unknown range will work out.. i guess good starting point would
> be
> >> check this out with simple 12-6 LJ case...
> >> ---
> >> sikandar
> >>
> >> On Wed, Mar 23, 2011 at 6:13 PM, Victor Ruehle <[email protected]>
> wrote:
> >>>
> >>> Hmm, don't see any clean way how to do this with the current
> >>> implementation. One might be able to subtract the known parts (which
> >>> can be calculated using rerun) using --force-trj and setting the
> >>> interval only to the unknown region. However this might lead to kinks
> >>> due to interpolation effects at the crossing.
> >>>
> >>> The clean way would require extending the code.
> >>>
> >>> 2011/3/15 Sikandar Mashayak <[email protected]>:
> >>> > Hey
> >>> > I am thinking of a problem where I have some physical understanding
> or
> >>> > constraint for CG-CG potential that I want to parameterize. So that,
> I
> >>> > know
> >>> > say from 0-r1 what is CG-CG potential values and want to determine
> >>> > values
> >>> > from r1-rcut.
> >>> > Is there any way I can set up such Force Matching calculations in
> >>> > Votca?
> >>> > thanks
> >>> > sikandar
> >>> >
> >>> > --
> >>> > You received this message because you are subscribed to the Google
> >>> > Groups
> >>> > "votca" group.
> >>> > To post to this group, send email to [email protected].
> >>> > To unsubscribe from this group, send email to
> >>> > [email protected].
> >>> > For more options, visit this group at
> >>> > http://groups.google.com/group/votca?hl=en.
> >>> >
> >>>
> >>> --
> >>> You received this message because you are subscribed to the Google
> Groups
> >>> "votca" group.
> >>> To post to this group, send email to [email protected].
> >>> To unsubscribe from this group, send email to
> >>> [email protected].
> >>> For more options, visit this group at
> >>> http://groups.google.com/group/votca?hl=en.
> >>>
> >>
> >
> > --
> > You received this message because you are subscribed to the Google Groups
> > "votca" group.
> > To post to this group, send email to [email protected].
> > To unsubscribe from this group, send email to
> > [email protected].
> > For more options, visit this group at
> > http://groups.google.com/group/votca?hl=en.
> >
>
> --
> You received this message because you are subscribed to the Google Groups
> "votca" group.
> To post to this group, send email to [email protected].
> To unsubscribe from this group, send email to
> [email protected].
> For more options, visit this group at
> http://groups.google.com/group/votca?hl=en.
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"votca" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/votca?hl=en.

Reply via email to