Re: problem with restraining a dihedral angle

From: Victor (ovchinnv_at_gmx.com)
Date: Tue Aug 13 2013 - 19:19:59 CDT

Sijia,

I do not know which version of the source code you use, but in the
versions I have worked from ( > 2008), the configuration you describe
would not work without lambda control. You can see this in the file
GlobalMasterFreeEnergy.C lines 39-87 (the update function). The
restraint manager only add forces if lambdaManager.getLambdas returns
true (there exists an active lambda object); in your case you do not
define any pmf or mcti blocks (lambda objects);

The solution would be to (1) modify the code to take the call to
addForces out of the GetLambdas conditional (a very simple mod), or (2)
to define a moving restraint using 'dihe pmf' instead of dihe, and add
an appropriate lambda controller, e.g. something like

pmf {
task=stop
time=xx
print=yy
lambda=0
}

Remember that the FErestraints code is not actively maintained, so, if
you find problems with it, you are basically on your own, and should be
willing to get your hands dirty. Of course, the third solution would
still be to use colvars, which replaced FErestraints (i.e. I do not
think you made it very clear why a harmonic vs. a cosine potential will
make a big difference for whatever you are trying to accomplish).

--Victor

On 08/12/2013 07:55 PM, Sijia (Scarlett) Dong wrote:
> Dear NAMD users,
>
> I'm using the method on this page:
> http://www.ks.uiuc.edu/Research/namd/2.6/ug/node34.html#SECTION000973
> to restrain one of the dihedral angles in a customized residue in my
> protein. The relevant lines in my configurations file is:
> -------------
> freeEnergy on
> freeEnergyConfig ret_urestraint
> -------------
>
> And ret_urestraint is a file with this content:
> -------------
> urestraint {
> dihe (MAIN, 296, C19) (MAIN, 296, C18) (MAIN, 296, C17) (MAIN, 296,
> C16) barr=10000 ref=-157.69
> }
> -------------
>
> Here is the problem: I have tried "barr" value from 0 to 10000 when I
> do minimization on the protein, but the dihedral angle I want to
> restraint is always around -161 or -162 deg which is not close to my
> reference angle. There is no much trend relating the barr value and
> the angle I'm getting except that when barr=0 it is -162 deg which is
> the same as not using the restraint. I'm using Amber force field.
>
> Does anyone know what I have done wrong to make the dihedral angle not
> getting closer to the reference angle? Are there other parameters I
> should be setting as well to make the restraint work? Is there a
> limitation related to this method such as it doesn't work with Amber
> force field?
>
> I'm not using the harmonic restraint here because I want to use a
> method in NAMD that has a equivalence to one of the LAMMPS restraining
> method to restrain the dihedral angle. The LAMMPS "fix spring" method
> doesn't work in the way the NAMD's harmonic restraint works, so I find
> this method which seems to be similar to LAMMPS' "fix restrain". If
> someone knows other methods that can achieve my purpose, I greatly
> appreciate it too!
>
> Thanks a lot!
> Scarlett
>
>

This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:23:34 CST