2024-11-25 Westbrook/Wagner Check-in meeting notes

General updates

• JW – Good work managing and helping with QC compute

  • BW – Thanks.

• Organometallic representation?

  • BW – Main toolkit warnings are

    • Radicals

    • …

    • JW – A lot depends on whether we’re implementing support for organometallics in OpenFF toolkit. If we ARE:

      • Figure out how to read molecules from cif (since it has formal charges and bond orders)

      • Make a branch of the OpenFF Toolkit where you try removing the radical check and see if it can support metals.

      • Use the QCSubmit pipeline as usual to do the submission and result retrieval

    • If we AREN’T:

      • Figure out how to read molecules from cif (since it has formal charges and bond orders)

      • Try making a pipeline for submission directly by going from RDKit → QCFractal (with CMILES)

      • Leave it as an exercise to the users to figure out how to ingest the dataset.

    • A tip for reading from CIF files into OFFTK is https://github.com/openforcefield/openff-toolkit/blob/main/utilities/make_substructure_dict/_cif_to_substructure_dict.py, which is probably more complicated than you need since it handles substructures (whereas you’ll only need to handle full molecules)

  • So immediate steps could be

    • Try submitting whatever you have to QCA to start teasing out QCFractal/Psi4 issues

    • Work on a pipeline to get from the conformer+bond order+formal charge info in the .cif file to an RDMol (this will be needed in either case above)

    • Figure out what the convention is for metal representations (especially formal charges + bond orders) in the cif file, and document what that is and that we’re using it.

• BW – Feedback on https://github.com/openforcefield/qca-dataset-submission/pull/412#pullrequestreview-2459578011?

  • JW (gave quick feedback but left LW as primary reviewer)

  •  

  •  

Trello

https://trello.com/b/dzvFZnv4/infrastructure