search

openforcefield / standards

A repository of the standards employed across the Open Force Field Consortium.
https://openforcefield.github.io/standards
MIT License
1 stars 3 forks source link

Description of `default_idivf="auto"` not in accordance with community understanding #60

Closed mattwthompson closed 3 months ago

mattwthompson commented 6 months ago

I interpret this

top-level attribute default_idivf, which can be an integer ... or "auto" if the barrier height should be divided by the number of torsions impinging on the central bond

I understood this to mean that, for example, if a single 6-term torsion is applied to the central bond in ethane, each idivfn should be 1/6. I appear to be completely wrong, though.

j-wags commented 6 months ago

Yes, this would be good to clarify. My understanding is that "auto" means "look at the molecular topology, count the number of unique 4-atoms paths that go through the central bond, and divide all the involved k's by that amount". So ethane's (CC) central bond would have idivf=auto evaluate to 9 for the central bond, while methylamine (CN) would evaluate to 6, and methanol (CO) to 3.

As far as I can tell, the current form of the SMIRNOFF spec does not identify which of our interpretations is correct.

lilyminium commented 5 months ago

FWIW Jeff's interpretation was my understanding of the phrasing as well, which as far as I can tell is also the implementation in AMBER.