Fast/ adaptive settings #1523

jthorton · 2025-09-22T10:37:09Z

This PR adds partial support for adaptive settings (#974) based on the transformation for the hybrid topology RBFE protocol. This is used to expose the fast settings for net neutral transformations. For these transformations the following settings will be updated:

nonbonded_cutoff = 0.9 nm
time_per_iteration = 2.5 ps
box_shape = "dodecahedron"
solvent_padding = 1 nm (complex) 1.5 nm (solvent)

Questions:

Are we happy with the CLI flag being adaptive settings?
Do we need to add more warnings about using fast settings?
Are we happy with how this is exposed at the API level do we need more guardrails?

I plan on adding tests once we are happy with the proposal.

Checklist

Added a news entry

Developers certificate of origin

I certify that this contribution is covered by the MIT License here and the Developer Certificate of Origin at https://developercertificate.org/.

IAlibay

Couple of quick thoughts - I'm not really into the flag (i.e. it doesn't feel like it adds anything useful), but it's not a strongly held view.

openfe/protocols/openmm_rfe/equil_rfe_methods.py

openfecli/parameters/misc.py

openfe/protocols/openmm_rfe/equil_rfe_methods.py

IAlibay · 2025-09-22T10:44:43Z

openfe/protocols/openmm_rfe/equil_rfe_methods.py

+                    f"{mapping.componentA.name} and {mapping.componentB.name}. "
+                    "A more expensive protocol with 22 lambda windows, sampled "
+                    "for 20 ns each, will be used here.")
+            warnings.warn(wmsg)


Thoughts on warning here but not warning on the default behaviour (below)? I.e. should this be just at INFO level since we expect it to be adaptive?

This was to maintain the current CLI experience, where this warning is displayed for charge change edges. However, more generally, perhaps an INFO-level log would be better for each edge, indicating if the settings are adapted from the defaults.

IAlibay · 2025-09-22T10:46:23Z

openfe/protocols/openmm_rfe/equil_rfe_methods.py

+
+        else:
+            # apply the fast settings we have benchmarked
+            protocol_settings.forcefield_settings.nonbonded_cutoff = 0.9 * unit.nanometer


Probably aim to only merge this PR after fast settings have been validated for net charge transformations - it would be good if we could avoid having to set any of these (or maybe we can set the default to the solvent case).

hannahbaumann

Thanks @jthorton this looks great! I agree with Irfan that this could also just be the new default without a flag, unless it would be too confusing for our users.

Other than that we'll probably just have to wait to see if we can switch charge changes to fast settings as well and tests.

IAlibay

Couple of things - we should just update the main settings objects so the changes are reflected everywhere.

IAlibay · 2025-10-17T12:56:00Z

openfe/protocols/openmm_rfe/equil_rfe_methods.py

                equilibration_length=1.0 * unit.nanosecond,
                production_length=5.0 * unit.nanosecond,
+                # fast settings
+                time_per_iteration=2.5 * unit.picosecond


Can you change this in the underlying settings file? That way it propagates to all the other Protocols: https://github.com/OpenFreeEnergy/openfe/blob/main/openfe/protocols/openmm_utils/omm_settings.py#L553

IAlibay · 2025-10-17T12:56:53Z

openfe/protocols/openmm_rfe/equil_rfe_methods.py

-            solvation_settings=OpenMMSolvationSettings(),
+            solvation_settings=OpenMMSolvationSettings(
+                # fast settings
+                box_shape="dodecahedron",


Same thing here, can you update this in the original settings object? that way we can update septop/abfes at the same time automatically

IAlibay · 2025-10-17T12:57:11Z

openfe/protocols/openmm_rfe/equil_rfe_methods.py

-            forcefield_settings=settings.OpenMMSystemGeneratorFFSettings(),
+            forcefield_settings=settings.OpenMMSystemGeneratorFFSettings(
+                # fast settings
+                nonbonded_cutoff=0.9 * unit.nanometer


Should we update this in gufe? @atravitz thoughts?

That would be great as it is the recommended cutoff for openff so we probably should use that everywhere.

updated in gufe

IAlibay · 2025-10-17T13:05:24Z

openfe/protocols/openmm_rfe/equil_rfe_methods.py

+                # fast settings
+                box_shape="dodecahedron",
+                # make the default padding work for all cases
+                solvent_padding=1.5 * unit.nanometer


I've raised #1589 for us to update settings in SepTop & AFEs.

This has broken some of the SetTop tests, is it best to update the solvent padding for SepTop here as well?

IAlibay

lgtm (putting aside needed changes in gufe)

openfe/protocols/openmm_rfe/equil_rfe_methods.py

openfe/protocols/openmm_septop/equil_septop_method.py

openfe/tests/protocols/openmm_septop/test_septop_protocol.py

atravitz · 2025-10-17T19:47:04Z

just a reminder to check that any changes are included on the CLI side, where applicable. #1591

codecov · 2025-10-20T10:48:22Z

Codecov Report

❌ Patch coverage is 98.36066% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 93.38%. Comparing base (3ebf7e3) to head (d3c4688).
⚠️ Report is 2 commits behind head on main.

Files with missing lines	Patch %	Lines
...ork_planner/relative_alchemical_network_planner.py	66.66%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1523      +/-   ##
==========================================
- Coverage   95.45%   93.38%   -2.08%     
==========================================
  Files         174      174              
  Lines       14553    14564      +11     
==========================================
- Hits        13892    13600     -292     
- Misses        661      964     +303

Flag	Coverage Δ
fast-tests	`93.38% <98.36%> (?)`
slow-tests	`?`

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

openfe/protocols/openmm_rfe/equil_rfe_methods.py

Co-authored-by: Irfan Alibay <IAlibay@users.noreply.github.com>

# Conflicts: # openfe/tests/protocols/openmm_ahfe/test_ahfe_protocol.py # openfe/tests/protocols/openmm_rfe/test_hybrid_top_protocol.py

atravitz

just an opinion on the news entry

news/rbfe_settings_update.rst

openfe/protocols/openmm_rfe/equil_rfe_methods.py

news/rbfe_settings_update.rst

Co-authored-by: Irfan Alibay <IAlibay@users.noreply.github.com>

github-actions · 2025-10-20T18:31:02Z

No API break detected ✅

jthorton added 3 commits September 22, 2025 11:05

add adaptive settings method

e876f7d

remove env file change

57f125b

expose the adaptive settings via the CLI

77c7b5b

jthorton marked this pull request as draft September 22, 2025 10:37

jthorton requested review from IAlibay, atravitz and hannahbaumann September 22, 2025 10:38

IAlibay reviewed Sep 22, 2025

View reviewed changes

fix indent, stop using frozen copy

b095603

jthorton self-assigned this Sep 22, 2025

jthorton and others added 3 commits September 25, 2025 10:53

Merge branch 'main' into fast_settings

54daf3a

Merge branch 'main' into fast_settings

d8f0f85

Merge branch 'main' into fast_settings

3771b89

hannahbaumann self-assigned this Oct 7, 2025

Merge branch 'main' into fast_settings

0dcc522

hannahbaumann reviewed Oct 7, 2025

View reviewed changes

jameseastwood added this to the v1.7.0 milestone Oct 7, 2025

IAlibay and others added 2 commits October 7, 2025 15:49

Merge branch 'main' into fast_settings

5271be4

Merge branch 'main' into fast_settings

d04a484

atravitz added the priority:high label Oct 14, 2025

atravitz and others added 7 commits October 16, 2025 08:59

Merge branch 'main' into fast_settings

daca91c

Merge branch 'main' into fast_settings

f7914b3

Merge branch 'main' into fast_settings

f49ba8d

pr feedback

3ec486b

typo

7701bf5

gj pycharm

6137bca

update tests

9d0498b

jthorton marked this pull request as ready for review October 17, 2025 12:51

IAlibay mentioned this pull request Oct 17, 2025

Update SepTop & ABFEs for fast settings #1589

Closed

IAlibay requested changes Oct 17, 2025

View reviewed changes

IAlibay approved these changes Oct 17, 2025

View reviewed changes

jthorton mentioned this pull request Oct 17, 2025

Update the default nonbonded cutoff OpenFreeEnergy/gufe#649

Closed

2 tasks

fix tests, add option to supply input settings

757a8fe

jthorton commented Oct 17, 2025

View reviewed changes

openfe/protocols/openmm_rfe/equil_rfe_methods.py Show resolved Hide resolved

jthorton commented Oct 17, 2025

View reviewed changes

openfe/protocols/openmm_septop/equil_septop_method.py Show resolved Hide resolved

jthorton commented Oct 17, 2025

View reviewed changes

openfe/tests/protocols/openmm_septop/test_septop_protocol.py Show resolved Hide resolved

jthorton commented Oct 17, 2025

View reviewed changes

openfe/tests/protocols/openmm_septop/test_septop_protocol.py Show resolved Hide resolved

jthorton commented Oct 17, 2025

View reviewed changes

openfe/tests/protocols/openmm_septop/test_septop_protocol.py Show resolved Hide resolved

jthorton and others added 2 commits October 17, 2025 16:29

Merge branch 'main' into fast_settings

c0c0d35

Merge branch 'main' into fast_settings

47e8f4d

atravitz mentioned this pull request Oct 17, 2025

check that all changes to defaults are reflected in the CLI documentation #1591

Open

hannahbaumann and others added 3 commits October 20, 2025 11:07

Merge branch 'main' into fast_settings

ff0358c

Merge branch 'main' into fast_settings

5a6355a

fix tests, remove redundant septop settings changes

052bf81

IAlibay reviewed Oct 20, 2025

View reviewed changes

openfe/protocols/openmm_rfe/equil_rfe_methods.py Outdated Show resolved Hide resolved

jthorton and others added 3 commits October 20, 2025 14:30

Update openfe/protocols/openmm_rfe/equil_rfe_methods.py

4be1093

Co-authored-by: Irfan Alibay <IAlibay@users.noreply.github.com>

fix nb tests, add news

2e64d95

Merge branch 'main' into fast_settings

bfe8d5a

# Conflicts: # openfe/tests/protocols/openmm_ahfe/test_ahfe_protocol.py # openfe/tests/protocols/openmm_rfe/test_hybrid_top_protocol.py

IAlibay approved these changes Oct 20, 2025

View reviewed changes

atravitz requested changes Oct 20, 2025

View reviewed changes

news/rbfe_settings_update.rst Outdated Show resolved Hide resolved

openfe/protocols/openmm_rfe/equil_rfe_methods.py Show resolved Hide resolved

IAlibay reviewed Oct 20, 2025

View reviewed changes

news/rbfe_settings_update.rst Show resolved Hide resolved

news/rbfe_settings_update.rst Outdated Show resolved Hide resolved

news/rbfe_settings_update.rst Outdated Show resolved Hide resolved

atravitz and others added 3 commits October 20, 2025 11:29

Update news/rbfe_settings_update.rst

3781e58

Co-authored-by: Irfan Alibay <IAlibay@users.noreply.github.com>

Apply suggestion from @IAlibay

971ed76

Co-authored-by: Irfan Alibay <IAlibay@users.noreply.github.com>

Apply suggestion from @IAlibay

d3c4688

Co-authored-by: Irfan Alibay <IAlibay@users.noreply.github.com>

atravitz approved these changes Oct 20, 2025

View reviewed changes

atravitz merged commit 774c42e into main Oct 20, 2025
10 of 12 checks passed

atravitz deleted the fast_settings branch October 20, 2025 18:32

Fast/ adaptive settings #1523

Fast/ adaptive settings #1523

Uh oh!

Conversation

jthorton commented Sep 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Developers certificate of origin

Uh oh!

IAlibay left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

hannahbaumann left a comment

Choose a reason for hiding this comment

Uh oh!

IAlibay left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

IAlibay left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

atravitz commented Oct 17, 2025

Uh oh!

codecov bot commented Oct 20, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

Uh oh!

atravitz left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

github-actions bot commented Oct 20, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

jthorton commented Sep 22, 2025 •

edited

Loading

IAlibay left a comment •

edited

Loading

codecov bot commented Oct 20, 2025 •

edited

Loading