[gmx-users] improper OPLS dihedrals in gromacs

[Mark Abraham]

On 20/10/2011 12:26 AM, mu xiaojia wrote:
> Dear GMX-users,
>
> I found a question in 2009 asking which format of improper does 
> OPLS dihedral gromacs use. I have the same question, is it periodic or 
> harmonic?

The quote below clearly says they are implemented as proper periodic 
dihedrals.

> If it is periodic, why in its ffbonded.itp file: #define 
> improper_O_C_X_Y 180.0 43.93200 2 what do they mean?

These defined macros are used in the .rtp file, as your quote says. See 
http://en.wikipedia.org/wiki/C_preprocessor#Macro_definition_and_expansion. 
The types of numbers and their order match that for dihedral type 1 or 
4, which are periodic. If they matched some other order then another 
dihedral type would be implied - though this is admittedly a bit 
perverse. See table 5.5. in the manual.

> Also, when I plug in tinker's improper dihedrals(which should be 
> harmonic) into gromacs, the results matches well whith each 
> other(tinker and gromacs). Does it suggest that gromacs uses a 
> periodic dihedral?

The quote you make below confirms this.

Mark

>
> Thanks, here is the part I quoted from ffbonded.itp
>
> [ dihedraltypes ]
> ; Improper OPLS dihedrals to keep groups planar.
> ; (OPLS doesnt use impropers for chiral atoms).
> ; Since these functions are periodic of the form 1-cos(2*x), they are 
> actually
> ; implemented as proper dihedrals [1+cos(2*x+180)] for the moment,
> ; to keep things compatible.
> ; The defines are used in ffoplsaa.rtp or directly in your .top file.
>
> --jia
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20111020/80564ac6/attachment.html>