[gmx-users] dihedral potential structure

[Mark Abraham]

Филипп Орехов wrote:
> Hello gromacs users,
> 
> I want to ask a simple thing, which nevertheless is not really clear to me,
> but it's sometimes too important. For instance, I have in my molecule 
> torsional
> angle around a double bond: -C-C(H)=C(H)-C-. The dihedral potential contains
> following four parts:
> -C-C=C-C-, a U1 from topology
> -H-C=C-C-, a U2=U3 from topology
> -C-C=C-H-, a U3=U2 from topology
> -H-C=C-H-, a U4 from topology
> Attention please, here the question is:  how  does the final potential 
> look like?  Is it just only a sum
> Utot=(U1+U2+U3+U4) or gromacs divides it on the number of elements of 
> the dihedral, I mean, in this case,
> Utot=0.25*(U1+U2+U3+U4)?

I'd guess they're just added. Set yourself up a simple example case to 
test: only define the dihedrals, and evaluate energies (1-step EM) with 
different dihedral constants.

Mark