[gmx-users] Question about redundancy in forcefield files and dihedral parameterization

[Eric Smoll]

Hello GROMACS users,

In the *types directives, what impact does redundancy have? Does a later
specification  override an earlier specification? For a concrete example,
consider a bondtypes entry:
[ bondtypes ]
A B 1 1.0 2.0
A B 1 1.0 3.0
Which specification will be used?

What impact does arrangement redundancy have on dihedral parameters? Will
the equivalent specifications shown below cause problems?
[ dihedraltypes ]
A B B C 3 1.0 2.0 3.0 4.0  5.0 6.0
C B B A 3 1.0 2.0 3.0 4.0  5.0 6.0
C B B A 3 1.0 2.0 3.0 4.0  5.0 600.0

When computing dihedral parameters of small symmetric molecules using ab
initio calculations, does one divide the weight or amplitude of the
LJ-subtracted dihedral potential across equivalent dihedrals along the same
bond? For example, does the computed dihedral spectrum of ethane (H-C-C-H)
get divided by 9?

Best,
Eric