2025.11.26 Cole Lab Check-In (Aus) Meeting Notes

NAGL MBIS Paper

DC: NAGL MBIS paper accepted. Discussion about data leakage? More than was expected.

LW: ASHGC didn’t use deepchem maxmin splitter so shouldn’t have the same issue.

DC: Maxmin splitter won’t differentiate if there are duplciates in the dataset.

LW: Mapped smiles can be different for the same molecule

Bespoke fitting progress

FC

• NM: Send over SMARTS matching example and we can take a look

Condensed phase fitting with SMEE and Descent

• LW: SMEE for nonbonded fits? Computational resources required?

• DC: JH only looked at a subset for fitting so can’t really say. Some changes in defaults in descent required.