2024-05-08 FF Fitting Meeting

Torsion multiplicity update

BW

• DM: so you’ve dramatically improved dataset coverage of specific chemistries – does looking at overall plots give good idea of these chemistries?

  • BW: good idea, will look into it

• JW: on RMSD – looks like the slope on the right hand side is improved a lot. Would prefer cumulative distribution plots

• JW: how are you feeling about idivf?

  • BW: not very confident at the moment

• LW: slightly off-topic, but we can’t currently charge SX3+1 because OE refuses to generate conformers for it. That being said it’ll assign charges if you give it another conformer, e.g. generated with RDKit