2025-04-16 FF fitting meeting

NAGL vdW refit

LW

LW will link slides here

• (slide 8) JW – Does this introduce a risk that some valence parameter splits+refits might become a blocker to our release of the nonbonded refit in 2.3.0?

  • LW – No, if the splits don’t work then the “v0” FF can be used which doesn’t have any changes

• PB – Is bromine where we see the most improvement with vsites?

  • LW – Br is actually the one where we see the vdW refit with vsites getting even worse. But we always did benchmarking after refitting vdW… though this may be different now since we have a more expansive phys prop benchmarking dataset now.

• DM – Unrelated - In last newcastle meeting, it sounded like smee is doing comparable force fields now. How do you see that fitting into things?

  • LW – Very excited about this. Most excited that their FF where they reset and expand out torsions does so well. I wonder about overlap with ex JHoeflich’s work. would love to try this after annual meeting.

  • DM – Yes, excited to be able to move away from legacy torsion methods, might get us “unstuck” in some ways.

  • LW – Yeah, just met with DC and FC last night, and they were looking at doing more automated typing. I pointed them to espaloma, but would love to put more effort into this in the future.

• LW – Also unrelated - I see there’s an alchemiscale release that might break pontibus.

  • JW – I’ll take care of this, and assume it’s my responsibility.

•  

•