2021-10-06 Industry benchmarks meeting notes

Owned by David Dotson, created
Last updated: Oct 18, 2021 by Lorenzo D'Amore

Participants

  • @Jeffrey Wagner

  • @Lorenzo D'Amore

  • @David Hahn

  • @David Dotson

Goals

  • Sage and OPLS submission update

  • Manuscript needs

  • Updates from team

  • Partner Season 1 Retrospective Survey

Discussion topics

Item

Presenter

Notes

Item

Presenter

Notes

Sage and OPLS submission update

Lorenzo

  • LD: same submissions in as last week; no new submissions yet

    • for OPLS using Janssen infrastructure for public molecules

    • have done almost 5,000 molecules out of 10,000

  • DD: appropriate to poke with another announcement

    • LD: will put together an announcement to prod folks

    • DD – Happy to help if you want reviews of drafts.

Manuscript needs

 

  • LD – Nothing here. Still awaiting results.

  • DD – May be good to phrase the messaging as “this is the last blocker to beginning work on the manuscript”. I know BSwope is working on it, may be good to contact him individually apart from mass email.

  • LD – I could ping BSwope, I’m in regular contact with him anyway.

Partner retrospective survey

David

  • DD: working to get this assembled later this week; will distribute to team for review

Updates from team

 

  • DH –

    • Nothing from me.

    • I just realized that I never updated the confluence protocol with the new OPLS instructions, we just emailed the corrected commands out to participants (todo from last week).

      • LD – Your solution worked for me, I’d be happy to update the protocol doc.

      • DH – That’d be great. The updated instructions+paragraph are in my email to the partners.

  • JW – no updates from me

  • DD – no updates. We have a protein-ligand benchmarks/OpenFF-OpenFE coordination meeting soon after this meeting which you all may want to attend.

  • LD – working to extend analysis of parameter violations, have many ideas from Gary, David H., David M.

    • focusing on 6 or 7 torsion violations

    • reducing this number is worth doing, so using fragmenter to fragment affected molecules

    • DD – gopt necessary?

    • JW – If it’s standalone then we should figure out what guarantee of support+stability we make. Otherwise we may tie it into the Toolkit/bespokefit as an example or CLI entrypoint.

    • Is anyone being blocked by the absence of gopt?

      • DD + LD – Not currently. XLucas

    • Options for moving forward

      • Make gopt a discrete package

      • Tie it into OFF Toolkit or bespokefit as CLI entrypoints

      • Do nothing / leave as openff-benchmark prototype

Action items

@Lorenzo D'Amore will update the protocol document with new OPLS merging instructions from email (taking over DH’s todo from last week’s meeting)
@Lorenzo D'Amore will poke partners for Sage RC2 results, perform individual outreach
@David Dotson will assemble partner survey on Season 1 retrospective
@David Dotson will lay out vision for openff-gopt, raise question in Slack if anyone actually needs that (right now, near future)

Decisions