[gmx-users] Problem with phenylglycine MD

[Ashutosh Jogalekar]

Hello
I am trying to run an MD on a short tripeptide contaning phenyglycine  
(PHG)- that is with the phenyl ring direct attached to the alpha  
carbon. I am using the OPLS force field. The peptide is (PHE-PHG-LEU)
I modified the .rtp and .hdb files accordingly. I put in the atoms,  
bonds, and improper dihedrals in the .rtp file. I put in the new name  
PHG in the aminoacids.dat file.
Everything seems to start ok, until I am done with the energy  
minimization. There is no error, and the minimization ends smoothly.  
However, at the end, the alpha carbon is distorted; it's not  
tetrahedral and the hydrogen bends towards the phenyl ring. Not  
suprisingly, in the rest of the simulation, the distortion is not  
relieved. I don't know what's going wrong, since there is no error as  
such and the .rtp and the .hdb files have been modified in an apt  
manner.
One thing is that there are no parameters for some of the dihedrals  
in the OPLS force field. So when I am doing grompp for the EM, the  
program naturally puts zeros in front of those dihedrals. Two  
examples are:
N  CA  CG CD1
HA CA CG CD1

Could this be causing the carbon to distort? If so, where can I get  
the parameters for these dihedrals?
Help will be really appreciated since I am stuck with this problem  
for days!
Thanks
Ashutosh Jogalekar
Emory University, Atlanta, GA
USA