Fwd: Local minima on minimiztion with colvars

From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Wed Jan 21 2015 - 13:01:33 CST

I forgot to mention that all three colvars include the index of all atoms
of the active center and the index of the six CA atoms at the protein
backbone to which the active center is bound. Does that create problems
with "centers" The active center is at one chain only of the four chains
of the enzyme.

fp
---------- Forwarded message ----------
From: Francesco Pietra <chiendarret_at_gmail.com>
Date: Wed, Jan 21, 2015 at 7:45 PM
Subject: Local minima on minimiztion with colvars
To: NAMD <namd-l_at_ks.uiuc.edu>

Hi:
I am at a system of a multichain enzyme with a bimetallic active center
bound through covalent carbon-carbon bonds. Amber parm10 data, parm7 file.
I rebuilt the system by removing all crystallization water and letting leap
to set the H-atoms, which removed some hard ckashes. The only warnings from
leap where about some short interatomic distances (1A the shortest)

Minimization with distance-rmsd-orientaion colvars. I got the impression
that minimization brings a portion of the active center (an acetate
bridging the two metals) into local minima, with some twisting. My question
is, could my colvar input be modified from run to run so that the issue of
local minima is overcome?

I should also point out that the minimization goes on if the harmonic only
comprises rmsd and orientaion colvars. If I include the distance colvars
"d", it crashes with "colvars: Error: in parsing "centers".

Following is the colvar input

# collective variable config file for "rmsd" and "orient" variable
# using group

colvar {
  name d

  width 0.2

  distance {
    group1 {

        atomNumbers 1628 2059...etc

      centerReference # use relative coordinates
      rotateReference # (translated and rotated frame of reference)
      refPositionsGroup { # Define frame of reference base on separate
atom group
        atomsFile PROT_BOX_refB.pdb # from separate file
        atomsCol B
      }
      refPositionsFile PROT_BOX_refB.pdb # initial coordinates for
reference group
    }
    group2 {
      dummyAtom (6.0, -1.0, -0.2) # fixed position
    }
  }
}

colvar {
  name RMSD # needed to identify the variable

  width 0.3
  rmsd {
     atoms {
        #atomNumbers 7893 7901 7902 7909 7910 7911 7915 7922 7930 7931 7938
7939 7940 7950 7951 7952 7955 7961 7962
                   # C2 O2 C5 O3 O4 C9 N1 C14 O6 C18 OE1
OE2 CB NE2 FE1 O7 FE O OXT
         atomNumbers 1628 2059 etc
  }

     refPositionsFile PROT_BOX.pdb
 }
}

colvar {
  name ORIENT

width 0.1

  orientation {
     atoms {

          atomNumbers 1628 2059 etc
  }

     refPositionsFile PROT_BOX.pdb
 }
}

harmonic { # define a moving harmonic restraint

  colvars RMSD ORIENT

  centers 0.0 (1, 0, 0, 0) # centered around zero RMSD and no rotation

  forceConstant 30.0 # unit is kcal/mol/[width]^2
 # where the width parameter is adapted to the dimensions of each colvar
}

This archive was generated by hypermail 2.1.6 : Tue Dec 27 2016 - 23:20:50 CST