(Complete) Protein-Ligand Benchmarks Automation via Folding@Home

Primary Driver

@David Dotson

Approvers

(OpenFF)

(Chodera Lab, and F@H, and ASAP)

(OpenFE)

Project approvers agreed to dissolve as goals were completed on 2023_10_30

Supporting Drivers

@Lorenzo D'Amore

Stakeholders

@Simon Boothroyd Irfan Alibay @Richard Gowers David Swenson @Jeffry Setiadi @Lorenzo D'Amore @Lily Wang Antonia Mey @David Mobley

 

Project Manager

 

Decision authority

Unanimity of Primary Driver and two Approvers (absences are vetos), only in synchronous meetings

Veto authority: Primary Driver, any Approver

Discussion/ Notification Venue

Weekly meetings (decision forum)

#free-energy-benchmarking channel on OpenFF slack (notification and discussion, no major decisions allowed here)

alchemiscale issue tracker (discussion venue)

Objective

Leverage Folding@Home for automated, continuous benchmarking of new generations of OpenFF forcefields via binding free energy calculations for a library of protein-ligand systems.

Due date

Oct 1, 2022 

Status

Complete

Problem Statement

Thus far in the project, protein-ligand benchmarks for OpenFF have been a manual process performed on Janssen cluster resources. As new generations of OpenFF forcefields are released, the burden of executing free energy calculations to assess improvement and deficiency will increase, as will the computational cost.

We would like to improve this situation by:

  1. Leveraging the immense computational resources provided by Folding@Home for near-continuous free energy benchmarking of new force fields against a library of protein-ligand systems.

  2. Introducing automation that minimizes human-in-the-loop requirements for this benchmarking into the future.

  3. Executing free energy calculations via both Gromacs and OpenMM.

 

Current project specification

(The link below should be considered an authoritative source of truth for this project at the moment. This may be updated in the future)

https://github.com/orgs/openforcefield/projects/1/views/1

https://docs.google.com/document/d/1zQ3FbpwKGzjg-AnMJr1qnnwADKno6Epk_8oAeqVEGEE/edit#heading=h.opo9vf5xelgn

 

Milestones and deadlines

Milestone

Deadline

Status

Completion

Milestone

Deadline

Status

Completion

Working group recruitment

“Read the google doc, provide use cases in the form of “I expect to put X in as an input, and get Y out as an output. State availability for weekly meetings starting as early as Feb 8”

Feb 4 2022

COMPLETE

@David Dotson
@Diego Nolasco (Deactivated)

First meeting - How project is managed, open discussion of use cases

Feb 15 2022

COMPLETE

 

Second meeting - Further refinement of use cases

Feb 22 2022

COMPLETE

 

Scope/use case review

Mar 8 2022

complete

@David Dotson
@John Chodera
@Jeffrey Wagner

Initial API defined (example notebook or command block for each user story?)

2022.09.15

COMPLETE

 

alchemiscale release 0.1.0

2022.09.15

complete

2023.04.17

protein-ligand-benchmark release 0.3.0

2022.09.15

IN progress

 

“MVP” - Core infrastructure stand-up as development environment to enable rapid iteration, involvement of stakeholders

2022.09.15

COMPLETE

2023.04.21

@David Dotson
@John Chodera
@Jeffrey Wagner

Software development, integration and testing cycles

2022.09.15

complete

@David Dotson
@John Chodera
@Jeffrey Wagner

Production environment stand-up and deployment; ideally using promotion mechanism from a QA environment

2022.10.01

COMPLETE

2023.04.23

@David Dotson
@John Chodera
@Jeffrey Wagner
@David Dotson
@John Chodera
@Jeffrey Wagner

Reference materials

  • https://openforcefield.atlassian.net/wiki/spaces/PS/pages/1806761985

  • https://docs.google.com/document/d/1wcTy4TqBIGbxdt9p2reVTY4J5iKAzF673DQsZeZ8vg8/edit?pli=1#

  • https://docs.google.com/document/d/10avT6qoE5tXjoCBvYDpP9vpI93FvhmgcAI9-0f-TElc/edit?pli=1#heading=h.y7qhjlqxse2m

  • https://drive.google.com/file/d/1_N3yFBRKU8HoE4XCSfPrY26UFlncVGA8/view?pli=1

  • https://drive.google.com/file/d/1XRpJ9OELd005BCZRn7uwEpMqu70_f8yh/view

Meetings