[gmx-users] Why is there a difference between an angle of 0 or 180 deg. for a type 9 proper dihedral with multiplicity of 2?

[Christopher Neale]

Dear Gromacs users:

I have been working on creating a topology for an exotic molecule. It contains aromatic rings and my parameters always seemed to allow the rings to buckle and become non-planer, much like a glucose ring would (though a little less extensively). However, I have managed to solve the problem by switching my proper (type 9) dihedral angles from angle = 0 degrees, multiplicity = 2 to angle = 180 degrees, multiplicity = 2. I thought that those two conditions should be equivalent and it was only by seriously simplifying the molecule down to a single ring and then toying with every conceivable parameter that I even hit on this. I am using gromacs 4.6.3 and have not tried other versions of gromacs, but this makes so little sense to me that I thought I would ask about it here. There are lots of proper dihedrals in the available force fields that use dihedrals with a set angle of zero degrees, though I do note that for any aromatic ring that I have seen they are always 180 deg (with multiplicity of 2 so that they can handle both cis and trans). Presumably I have just missed some obvious definition, but at least I can verify that if I switch even one proper dihedral from angle = 180 back to angle = 0 (with multiplicity = 2 in each case), then I start to see deformation of the ring's planarity.

Thank you for your help,
Chris.