2022-08-02 Protein-ligand benchmarks meeting notes

fah-alchemy - current board status

David Dotson

• fah-alchemy : Phase 1 - MVP

• 4 weeks out from 9/1 deadline for biopolymer benchmarking

• focused on resolving approach to tokenization and serialization used within gufe via #39 and #36

  • DD – Current focus is on GUFE #39, about serialization and tokenization.

  • DS – Mostly now about writing unit tests

  • DD – Are we OK with the DAG system serializing things in a DASK-like way? Can we use DASK core to handle tokenization of protocol units/can we use the same the system as we’ve used for other GUFE objects for protocol units?

    • DS – I think we can use the same system as we’ve been using. It puts a bit more work on the writer but that’s fine.

    • DD – I have more thoughts on this but let’s revisit it next week.

• DD – starting up MVP of Executor, Scheduler, and ResultServer for FAH this week

• DD – I was out the past week, will be out the coming week. So not much has changed on my tasks.

ProtocolSettings taxonomy update

Mike Henry

• MH – Met with RG and DD last week, and tried to identify points of ambiguity/identidying which parts of the FF we need and which we don’t need. So, which parts we have inside of this and which parts can be external. We can talk about this in more depth later, I’d like to get things a bit more thought out before we discuss.

• JW – Refresh my memory?

  • MH – Ensuring that we store things like cutoff info from force fields (and other global setting/nonbonded stuff.) But not per-particle parameters.

protein-ligand-benchmark - 0.3.0 update

David Dotson

• IA – Caught up with MBoby last wednesday, decided which systems to fix and how to fix them

• JC – Re-prepping with new assay pH and exptl conditions?

  • IA – We decided to handle assay conditions in 0.4, not 0.3. The big push here was to prepare using schrodinger protein prep’s command line tool to record all args.

  • IA – Some docking RMSDs were really high, it may be that the docking constraints weren’t working properly. We’ll try to come back to this either for the 0.3 or 0.4 release.

• IA will message MBoby for status update today

• JC will meet with MBoby tomorrow

• IP – I raised some issues about SDF file parsing, and was told to escalate to RDKit issue tracker. On the RDKit tracker, they said that the SDF with no newline was non-format compliant.

  • IA – I think that the no-newline thing is valid.

  • IP – Link to RDKit issue where GLandrum says that the newline is a requirement in the spec:

    • Uses biovia spec http://help.accelrysonline.com/ulm/onelab/1.0/content/ulm_pdfs/direct/reference/ctfileformats2016.pdf (page 75)

  • DD – I may have made this mistake initially by making SDFs with no blank line.

  • IA – I’m going to review my sources and may contest GLandrum’s answer - Let me get back to

• JC – Any other issues with file formatting, IP?

  • IP – Nothing else

  • IA – There was a mapping issue, which I raised in perses. Basically we’re seeing a problem with hybrid topologies.

risks to project, timeline?

David Dotson