Page Comparison

Wrapping up AM1/restraint studies

• CD posted on GH issue with more details about his his scripts check for connectivity rearrangements and do

• optimization ambertools AM1

  • Github link macro
    link https://github.com/openforcefield/openff-toolkit/issues/1170

• Data storage?

  • CD – Important figures are in notebook.

  • OM – So, imporant figures should be saved, but I don’t think we need to save all the calculation outputs

  • JW – Random conformer vs. am1 conformer plot notebooks?

    • CD – I can add that notebook as well.

  • JW – If the raw data/scripts are in one folder, or can be thrown into a folder and tarred to under 10GB, I think the best way forward would be to make a release of the above repo and just toss the tarball up as a release asset. It’s fine if it’s messy, don’t spend a lot of time cleaning it up. If needed you could just use the releasenotes to say “there’s lots of experiments in there, most of them are garbage, you’re probably looking for directory X, Y, and Z”

  • CD – For crediting people like SB, where is the appropriate place for that?

    • JW + OM – Either releasenotes of top-level repo README. README is probably slightly better.

Update on ELF studies

• CD – Last time, we asked whether the number of INITIAL conformers form each method would affect the agreement when we selected the ELF conformer.

Image Added

Image Added

Caveat - These ensembles were generated using a 0 RMS cutoff, so it’s possible that ELF selection is disregarding closely-spaced conformers before doing any energy evaluation.

CD – Note that the ELF selection even in the big cases only took 2 or 3 secs, so the dominant cost for those was conformer generation, which was taking 20-60+ seconds.

• (General) – In all cases, we should have AT LEAST 5 x n_atoms, but it would be OK to have a hard cap at 200 or 300

Next steps for CD projects

• JW – We have working session with Vanderbilt after this, on mixed material/nanoparticle and OpenFF organic solvent simulaitons. Could be a good way to see the state of the science/infrastructure.

• Next steps:

  • Pull in Shirts to help advise on future projects:

    • Making a foyer-powered workflow to look at an interesting material/chemical question

    • Docking rescoring (either total or just strain energies) using OpenFF.

    • Graph charge/post translational modification testing

    • Alternate functional forms in the FF (would OM have to run the calcs on Lilac?)

    • Reoptimizing BCCs against ESPs (partially started by SB and OM, progress stopped but datasets may be ready)