2020-10-12 Sprint Planning Meeting notes

Date

Oct 12, 2020

Participants

  • @Jeffrey Wagner

  • @David Hahn

  • @Simon Boothroyd

  • @Matt Thompson

  • @David Dotson

  • @Pavan Behara

Discussion topics

Notes

Notes

Roundtable updates

  • DH – Working on manuscript / analysis of benchmark sets. Started reformatting analysis in notebooks, and uploaded to GH. Binder-ized the same. Worked on MT’s review of PLBenchmark repo – Ready for re-review.

  • SB – Refactored evaluator, using thermodynamic fluctuation formula for gradients, something else. Working on manuscripts.

    • DD – Touch base about scaling out evaluator / drafting proposed approach?

    • SB – Pretty tied up the next few weeks, and internet has been spotty.

    • DD – I’ll start drafting my thoughts on confluence and will touch base with you when you’re more available.

  • MT – Mostly just continuing to iterate on some PRs. Attended RDKit UGM. Writing up system spec.

    • JW – Anything useful out of UGM?

    • MT – Ntohing too surprising. GL gave a nice “state of the rdkit” talk, talked about funding models. Mentioned OpenFF’s model WRT sustainability.

  • DD –

    • Submitted protein fragments torsiondrives – Required figuring out which torsions to drive/which structures to use. Chugging through torsiondrives.

    • Prepared datasets for Maat and Rufa, waiting on DR to make dataset adhere to standards. Maat is working to add CMILES on submitted molecules.

    • PRP set up infrastructure for autoscalers (!!!!!!)

    • Benchmarking call coming up this week. JH is preparing some scripts, I’m also preparing a big-picture plan. Will ask DH and JH for review before we discuss.

      • DH – GT set up a call with pharma partners on the 23rd. I’ll be presenting there, will include some content from DD’s presentation.

      • DD – Anything in particular I can prepare for this?

      • DH – Not sure. I’ll let you know later as I’m drafting.

    • JW – Hypernetlabs will provide a one-pager about security details and such of running on Galileo. I’ll ask for them to have this ready for the 23rd.

  • JW –

    • Worked with MT on WBO bonds

    • Worked on N-1 CIMH

    • Conda-forge

      • Packaging of OpenMM is moving forward nicely

      • MT – Are theere significant reasons we couldn’t cut this on omnia? If not, then c-f is an unnecessary blocker to us making a py38 release.

      • JW – Omnia has been broken for some time,not sure if this would affect openmm. Omnia-dev is a dangerous place to build. PE has largely decided to delay the 7.5.0 release until the c-f migration is done, not sure about his whole reasoning.

      • MT – This is dangerous, because our entire stack is prevented from moving to py38 until the OMM migration moves forward.

      • DD – I’m happy with the current plan, because PE is very motivated to get this big migration done once and for all.

  • PB –

    • Finished PR for Topology.to_file(format='pdb', coords)

    • Specced out project for reliability of loading PDB/molecules based on limited info

    • Working with JM on torsion barrier height calculation and WBO harvesting. Will reach out to DD and TG for review on those.

Sprint planning

  • 0.8.0 release?

    • JW – Would like to see if I can finish N-1 CIMH before cutting. Should be doabl before Wednesday

  • Sprint planning – New ticket: Nonbonded refactor per issues opened by SB

  • Sprint planning – New ticket: SMARTS matching performance WRT protein FF

  • Sprint planning – New ticket: Review WBO torsion/barrier height PR for PB

    • DD – Will need to prep for pharma benchmarking meeting. Will probably have availaiblity later in the week.

  • Sprint planning – New ticket: (re)review of PLBenchmarks

    • MT – Can take care of second iteration here

Action items

Decisions