2024-06-17 Mitchell/Wagner Check-in meeting notes

General updates

• JM –

  • I’m going to start my holiday early and would like to take this tues+weds off in addition to what I’d requested before. So I’ll be offline the rest of this week and all of next

    • JW – Sounds good, please just update JustWorks.

  •  

• JW –

  • BWestbrook is working on getting QCSubmit examples updated+tested. Should we incorporate into openff-docs after that? They’re pretty text-y and I’m not sure that we want to advocate people using it directly.

    • JM – It’d be good to have QC examples on the docs page, though they may need some TLC. The QCSubmit vignette about fetching from QCArchive was pretty sharp. So I’d be in favor of folding them in.

    • JW – Good idea. I’ll pass this on to BW.

  • Nice work getting the mini-workshops up!

  • I’ve drafted a slack #general blast for the mini-workshops (JW will share via slack DM)

    • JM – Looks good

    • JW – I’ll make the announcement so that people aren’t pinging you while you’re out.

    • (JW posted announcement)

•  

Trello

https://trello.com/b/dzvFZnv4/infrastructure

PDB loading review

• JM – The whole PDB is 43GB (this is small). I can load it into the PDBFile classes in about one hour. So one hurdle might be runtime - It might take a lot of time to load into OpenFF Mols. So that might be a place to look at performance improvements. Also not sure how long PDBFixer processing will take.

• The process will look like “get an inventory of the whole PDB, according to what we can/can’t load”, and this might involve subsampling or something like that. So one prong of the effort is getting an inventory of the whole PDB, and the other prong is loading things. The former would be classifying/describing inputs by things like

  • Do they have any non-CCD residues

  • Do they have metals

  • Do they have Hs

• (JM and JW sketch out whiteboard below, update next series of to-dos on trello)

•