Page Comparison

Participants

...

Recording: https://drive.google.com/file/d/1_s2Arh7pzdcf9I3srHbQsj5-QOUdayv1/view?usp=share_link

Discussion topics

...

Item

...

Presenter

...

Presenter
Notes
JW
In person meeting schedule: https://docs.google.com/document/d/1ejGdhhipkIZJOa0mmV5IjTREbxcYGw1ywKryDh2ALdc/edit#
Let me know ASAP if there’s anything glaringly wrong - Will send for initial posting in about 24 hours (though we can make small changes after that)
Retreat schedule is coming together - Suggestions for activities?
Must be unrelated to deliverables and come with no implication of followup
Current suggestions include a hike to the beach, seeing if ChatGPT can eat our docs and run valid simulations, or a discussion where Josh M nerds out about his opinions on docs.
JW
Bespokefit 0.2.1 released - Compatible with ForceBalance 1.9.5 and OFFTK >=0.11
Interchange 0.3 just dropped - If your plugin broke, check out the docs and or contact matt
Expect OFFTK 0.13 soon - Will come with ability to load multicomponent PDB files (proteins+waters+monoatomic ions+small mols) to Topology objects
PB
Sage 2.1.0 update
Slides - https://docs.google.com/presentation/d/1dQC47CjLbm67rrz-ESAFOlz_RtKJ3Pp4PVrN0Wgmmow/edit?usp=sharing
MG – Slide 13 - Was talking with CSimmerling earlier and he mentioned fitting to differences - that is, “is the MM energy difference similar to the QM energy difference?”
PB – BSwope had mentioned analysis like this. Similar to kendall tau maybe?
MG – Not sure about the relation to kendall tau, but fitting directly to ddE might help
PB – Yeah, and another thought from trevor was to do single point energies and compare (that is, skip optimization so the geometries are exactly the same)
TG – To be clear – It sounds like what CSimmerling was doing was the same as what we’re doing, but not optimizing?
MG – I may not be remembering this right, but his point was that everything we’re doing now is looking at the difference of everything to one conformer, whereas they do all-to-all differences.
CC – Yes, that’s correct, they fit to all-to-all differences, and they also don’t do optimization.
TG – Interesting. I wonder if this manifests as our sharp peak at 0 (slide 13).
DM – Yeah. this overall looks better in most metrics but ddE is muddled. But overall this looks like a good improvement.
DM – One big question is “what do we want to see before we release this?” MOsato from my group can run FE calcs, also Ben Cossins from Exscientia mentioned that they can test this out in their workflow.
PB – I’ll send this RC to Ben.
PB – As long as it’s a release candidate, is it ok to put it in the repo?
JW – Yeah, let’s work together on this
.
DM
As of this morning, Project Manager candidate James Eastwood is accepting the offer, and we’ll set start date (probably after May 18).

Versions Compared

Old Version 1

New Version Current

Key

Participants

Discussion topics

Action items

Decisions

Presenter	Notes
JW	In person meeting schedule: https://docs.google.com/document/d/1ejGdhhipkIZJOa0mmV5IjTREbxcYGw1ywKryDh2ALdc/edit# Let me know ASAP if there’s anything glaringly wrong - Will send for initial posting in about 24 hours (though we can make small changes after that) Retreat schedule is coming together - Suggestions for activities? Must be unrelated to deliverables and come with no implication of followup Current suggestions include a hike to the beach, seeing if ChatGPT can eat our docs and run valid simulations, or a discussion where Josh M nerds out about his opinions on docs.
JW	Bespokefit 0.2.1 released - Compatible with ForceBalance 1.9.5 and OFFTK >=0.11 Interchange 0.3 just dropped - If your plugin broke, check out the docs and or contact matt Expect OFFTK 0.13 soon - Will come with ability to load multicomponent PDB files (proteins+waters+monoatomic ions+small mols) to Topology objects
PB	Sage 2.1.0 update Slides - https://docs.google.com/presentation/d/1dQC47CjLbm67rrz-ESAFOlz_RtKJ3Pp4PVrN0Wgmmow/edit?usp=sharing MG – Slide 13 - Was talking with CSimmerling earlier and he mentioned fitting to differences - that is, “is the MM energy difference similar to the QM energy difference?” PB – BSwope had mentioned analysis like this. Similar to kendall tau maybe? MG – Not sure about the relation to kendall tau, but fitting directly to ddE might help PB – Yeah, and another thought from trevor was to do single point energies and compare (that is, skip optimization so the geometries are exactly the same) TG – To be clear – It sounds like what CSimmerling was doing was the same as what we’re doing, but not optimizing? MG – I may not be remembering this right, but his point was that everything we’re doing now is looking at the difference of everything to one conformer, whereas they do all-to-all differences. CC – Yes, that’s correct, they fit to all-to-all differences, and they also don’t do optimization. TG – Interesting. I wonder if this manifests as our sharp peak at 0 (slide 13). DM – Yeah. this overall looks better in most metrics but ddE is muddled. But overall this looks like a good improvement. DM – One big question is “what do we want to see before we release this?” MOsato from my group can run FE calcs, also Ben Cossins from Exscientia mentioned that they can test this out in their workflow. PB – I’ll send this RC to Ben. PB – As long as it’s a release candidate, is it ok to put it in the repo? JW – Yeah, let’s work together on this .
DM	As of this morning, Project Manager candidate James Eastwood is accepting the offer, and we’ll set start date (probably after May 18).