[gmx-users] negative dihedral cp (during cgenff conversion)

[Mark Abraham]

On 15/08/2011 10:50 AM, Peter C. Lai wrote:
> Random noob question:
> Should the cp field (field 7) in [ dihedraltypes ] always be positive or can
> it be negative?
>
> In the CGenFF conversion, the CHARMM dihedral has:
>
> OG302  CG2O2  CG321  CG331     -0.1500  1   180.00 ! LIPID methyl proprionate
> when I convert to gromacs:
>
> [ dihedraltypes ]
> ; i     j       k       l       func    phi0    cp      mult
> OG302  CG2O2  CG321  CG331      9       180.0   -0.62760    1 ;LIPID methyl prop
>
> This is the only instance I have seen of a negative number for the dihedral
> term, so I wonder what the right way to handle it is...
> (I suppose I could just run it and see...)

There are a small minority of CHARMM27 dihedrals that do this. There's 
no intrinsic reason why the potential should be positive or negative. 
Forces and energy differences are unaffected by the sign of any 
component of the energy. Probably they result from optimization of these 
parameters to reproduce reference quantites in the context of other 
fixed parameters.

Mark