2023-08-23 FF Fitting Meeting

Strike team issues

BWestbrook

Open

• t129

  • TG – what’s the periodicity?

    • LW – 1

  • DLM – we might not need too much data for in-ring torsions since they’re not often rotated

• t164

  • DLM – could look at database of molecules. I suggest avoiding eMolecules since a lot of the structures can be wrong

QM Internal coordinate energy

TGokey

https://docs.google.com/presentation/d/1VpxSxgsfcHrmzqm3JKos7qH2JZrCEn2TbLhORMq1XX0/edit?usp=sharing

• Slide 7:

  • DLM: what’s the difference between these two slides?

  • TG: left graph is G, right graph is inverse, which is used in transformations. We build the left hand side but use the right hand side.

• BS: how are ICs generated?

  • TG: I use geomeTRIC, then I add NB coordinates by treating them as a pair

  • BS: Would you use 1-5 internal distances?

  • TG: Yes, they’d be in the NB part. We want to account for that and project out of the bonds/angles. 1-4s also.

• BS: might you need to account for stretch-bond or angle-bend coupling?

  • TG: yeah that’d be interesting, I could zero the black squares on slide 7

• PB: slide 16 – aren’t the right hand axes very low?

  • TG: Yes – it’s only for the 4-5 bond, and you could interpret it as the bond doesn’t need to move much.

• LPW: it is possible to completely represent all possible changes in a molecule using interatomic distances. My suggestion is that you try a Gram-Schmidt orthogonalisation

  •