[gmx-users] Is the discontinuity in type 2 improper dihedrals at 90 deg. or 180 deg. ?

[Christopher Neale]

Dear Justin:

Thank you for the insight. I had mistakenly thought that the combination of two overlapping impropers was what was inhibiting omega inversions. Obviously you are correct. 

Thank you,
Chris.

On 12/13/14 9:28 PM, Christopher Neale wrote:
> Dear Users:
>
> In gromacs, the charmm27 force field uses type 2 (non-periodic) improper dihedrals to maintain planarity of peptide bonds.
>
> Type 2 dihedrals are described on page 70 of the 4.6.7 manual and page 80 of the 5.0.2 manual, which contain the line (identical text):
> "Since the potential is harmonic it is discontinuous, but since the discontinuity is chosen at 180 deg. distance from xi_0 this will never cause problems."
>
> I'm a little confused about improper dihedrals, so maybe I am just not getting it, but would not the discontinuity be at +/-90 deg. (rather than 180 deg.) ? If the discontinuity were at 180 deg., then a type 2 dihedral should be able to discern between cis and trans peptide bonds, which I believe that it can not do (my interest here is eventually being able to discriminate against cis peptide bonds at very high temperatures).
>

Isn't this an issue of proper dihedrals, not impropers?  A cis-trans 
isomerization involves an actual bond rotation about C-N; that's not what the 
impropers are acting on.  If this were related to improper terms, you'd have 
C-N-CA in the same plane, but the H bending over backwards into something 
totally non-physical.

> Also, for type 2 improper dihedrals, isn't it the differential that is discontinuous, not the potential?
>

The nature of the discontinuity is described in 
http://www.charmm.org/documentation/c39b1/parmfile.html

-Justin