Ligpargen produces Torsions with all k=0.00000 in OpenMM .xml file

[bkpgh]

Using the web server on multicyclic molecules, the OpenMM .xml file contains some torsions that have all of the force constants = 0.00000.

A simple example is DABCO (smiles=C1CN2CCN1CC2) for which the output contains lines including:

Improper... class1="C807" class2="N802" class3="H819" class4="C806" k1="0.000000" k2="0.000000" k3="0.000000" k4="0.000000" ...

For some other molecules, both Proper and Improper torsions like this appear.

Are these non-functional torsions evidence of

1. a bug: there should be non-zero k's in there
2. torsions that should have parameters, but they just don't exist in OPLS-AA
3. torsions that resulted from a search, don't exist in OPLS-AA, aren't needed, and they can be deleted.
4. Something else?