[gmx-users] error in OPLS-AA dihedral parameters

[Malcolm Gillies]

On Tue, 2003-07-22 at 17:04, I wrote: 
> ffoplsaabon.itp contains the following dihedral parameter
> 
>   H      N2     CA     NC      3      2.93076  -2.93076   0.00000  
> 0.00000   0.00000   0.00000 ; Adenine RZ
> 
> this applies to the N1-C6-N6-HN6(1/2) dihedral in adenine.

On further investigation, this appears to be a typographical error in
the original OPLS-AA parameter files, which has been automatically
translated into the GROMACS ffoplsaabon.itp file.

I've found what appears to be the source of the parameter in question.
The relevant lines from the Schrodinger IMPACT-format OPLS-AA
paramstd.dat file (which translates directly to the line given above):

NC-CA-N2-H        2    1.400    0.0      1.0      Adenine RZ

[
compare this to (elsewhere in the same file):

CB-CA-N2-H       2    1.400  180.0    2.0    adenine RZhou
]

A copy of the file "OPLS All-Atom Parameters for Organic Molecules,
Ions, & Nucleic Acids  5/01", provided by Bill Jorgensen, has the
following entries:

[     V1        V2        V3        V4  (kJ/mol)              ]

215   0.0       2.03      0.0       0.0        ??-CA-N2-H     
aniline-like
215   0.0       2.03      0.0       0.0        N?-CA-N2-H     
aniline-like
215   0.0       2.03      0.0       0.0        CB-CA-N2-H     
aniline-like

so the correct solution would appear to be to delete the erroneous line
and let the torsion pick up the wildcard parameter in ffoplsaabon.itp:

  H      N2     CA     X       3      8.49920   0.00000  -8.49920  
0.00000   0.00000   0.00000 ; aniline-like

cheers,

Malcolm
--
Malcolm Gillies <Malcolm.B.Gillies at anu.edu.au>
Postdoctoral Fellow, Computational Proteomics and Therapy Design Group,
John Curtin School of Medical Research, Australian National University