2024-02-07 Mitchell/Wagner Check-in meeting notes

General updates

• JW – Use clockify. There’s a whole process for this, please complete ASAP.

• JW – New OFFTK and Interchange releases out with support for from_openmm, we ensured that it will handle your RNA notebook. Though the vdW energies are a little weird, so warn about that.

  • JM – I’m generally seeing funny numbers from different engines on vdW.

  • JW – Please let MT know.

  • JM – And incorrect bond energies for LAMMPS.

  • JW – MT mentioned that constraints aren’t being passed to LAMMPS correctly.

• JW – I’ve cleared up most of my backlog so I’m stealing some of your “ready to merge” items on Trello

• JW – We’re getting an awesome turnout of serious looking people for the miniworkshops!

Workshop planning

https://openforcefield.atlassian.net/wiki/spaces/COMMS/pages/2719055920

• JM – I’ve got a story for the SMIRNOFF workshop that I’m happy with. Found a PDB with hexanoic acid, which is a great example of SMIRNOFF parameter hierarchy use. Also found one with NADP. Can show energy comparison with OpenMM and GROMACS (we’re getting different energies for AMBER and LAMMPS). Will be good example of AMBER protein FF + Sage. Could also show off bespokefit.

• JM – The RNA thing works. Need to decide whether to packmol it or how else to come up with initial coords.

• JM – Protein prep is already pretty much together, just need to figure out docking.

• JW – Event invites. If you want to you can do it, but otherwise I or JEastwood can take care of it.

  • JM – I’d prefer to focus on notebooks.

  • JW – Sounds good, I or JE will take care of it.

• JM – Bespokefit NOT compatible with new QCF, but new toolkit ONLY compatible with new QCF

  • (discussion)

  • JW – Unfortunately the best way forward here is to use a different environment for bespokefit.

  • JM – There’s still an issue where bespokefit 0.2.2 and earlier can install with QCPortal >=0.50

    • JW – I’ll patch this

• JM – Also found an issue with interchange unit conversion from openmm. Communicated directly with MT about it, seems to be rapidly approaching resolution.

• JM – Any vignettes you really want (or want to keep around?)

  • JW – The RDKit reaction one is awesome.

  • JM – Tried making a vignette on partial tempering, a replice exchange style thing where you scale energy terms for part of the system. In theory lets you heat up solute without heating up solvent. Found a bunch of interchange trouble wiht this, worth digging into?

    • JW – I don’t think so, this doesn’t so much sound like it emphasizes unique stuff you can do wiht OpenFF since the same workflow could just as easily go in OpenMM. openff-toolkit 0.15.2 pyhd8ed1ab_0 conda-forge

  • JM – NAGL example?

    • JW – Yes, could be good to label it “yellow” and do a comparison of assigned charges to something established like antechamber

  • Docking rescoring?

    • JM – Could just make protein interchange once and then swap in different docked ligands

    • JW – This might be getting too complicated, then we’re using interchange + operator, and mixing AMBER and SMIRNOFF FFs…

    • JM – Could use GBSA?

    • JW – Something less complex, for example a loop that mutates a SMILES to get a mol that has a certain shape (for example, the digit 4) could be nice and snappy.

    • (General) – Or linker design between two ok-binding fragments.

      • JM – This sounds strictly more complicated than docking rescoring.

  • JW – Also, we should keep the NCAA vignette

    • JM – Should we update it to use NAGL?

    • JW – Up to your judgement.

    •  

    •  

    •  

    •  

To dos

1. (high) Do clockify

2. (high) Prepare workshops

3. (medium) Fix openff-docs CI/figure out what’s up with deployments

4. (low) Propagate PDB-bond loading options into nglview