Dear All,
I’m working on a system consisting of a DNA helix with 26 residues. I would like to calculate the free energy difference between two conformations. To calculate the entropy contribution to the free energy I have generated snapshots from the md trajectory. I have then energy minimized the snapshots with sander: 500 steps of steepest descent and 45500 steps of conjugated gradient, input file as below:
#########################################
Sander minimization, steepest descent-conjugated gradient, distance dependent dielectric constant
&cntrl
imin = 1,
ntmin =1,
maxcyc = 50000,
ncyc = 500,
ntb = 0,
ntc = 1,
ntr = 0,
cut = 100.0,
dielc = 4.0,
/
&ewald
eedmeth=5,
/
#########################################
After the sander minimization I have attempted Newton-Raphson minimization with nmode. The input file is pasted below (this was only for one minimization step to produce the output file, several steps have also been performed):
#########################################
Newton-Raphson minimization, nmode
&data
ntrun = 4,
maxcyc = 1,
scee = 1.2,
dielc = 4.0
&end
END
#########################################
Unfortunately the output file shows multiple dihedral angle errors:
#########################################
DIHEDRAL ANGLE ERROR: 1 90 93 102 105
DIHEDRAL ANGLE ERROR: 2 90 93 102 105
DIHEDRAL ANGLE ERROR: 3 90 93 102 105
DIHEDRAL ANGLE ERROR: 62 3 6 15 75
DIHEDRAL ANGLE ERROR: 63 3 6 15 75
etc.
#########################################
but the calculation continues.
I can see this error is written in ephi.f. But my knowledge of fortran isn’t good enough to track down exactly what’s wrong here. I thought there might be an issue with reading the topology file and tried using an old-format topology file but got an identical error.
The values for the energy terms calculated by nmode are identical to the final energy terms from the sander minimization which puzzles me at least for the dihedral term if there’s a problem reading dihedrals properly.
Has anyone encountered this error? Is it an innocuous warning?
Best regards,
Michael
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Nov 22 2010 - 11:00:04 PST
