Open Force Field Infrastructure Roadmap 2020

List of infrastructure aims and tasks for 2020. Last updated on Sept 24, 2020.

Labels

Category

Labels

Category

Labels

Priority

high | MEDIUM | LOW

Effort

high | MEDIUM | LOW

Status

Not started | In Progress | PROTOTYPE | Completed | BLOCKED |

Roadmap

Infrastructure tasks

Priority

Effort

Blocking science?

Infrastructure Dependencies

Start date

End/Due date

Status

Driver

Architecture / General infrastructure

Package namespace reorganization

Low

 

 

Likely not migrating toolkit and forcefields repo until toolkit 1.0, but new packages should follow the openff namespace convention

 

 

IN PROGRESS

@Matt Thompson

95%+ core package uptime and deployment

(OpenFF TK, s99F, OpenFFs)

High

High

 

Will be made easier by conda-forge migration once OpenMM gets moved over

Ongoing

 

IN PROGRESS

@John Chodera @Jeffrey Wagner @Matt Thompson @David Dotson

Monitoring dashboard, Nightly builds, deployment tests, error severity/triage policy

medium

 

 

Nightly builds will be easier after conda-forge migration

April 2020

 

IN PROGRESS

@Jaime Rodríguez-Guerra @Matt Thompson @David Dotson

Automated QC mol submission

High

 

No, but slowing some things/diverting effort from elsewhere.

Chemper package creation

Migration of CMILES and Fragmenter functionality into OFFTK

May 2020

 

IN PROGRESS PROTOTYPE

@Joshua Horton @David Dotson

Automate finding FF discrepancies and submitting torsion drives – find more molecules that have underexplored parameters, eats large molecule datasets and compared OpenFF to ANI energies, and flags most discrepant for QM calculation. Emphasize plugin architecture so other people can easily add new criteria. Relevant slack discussion.

medium

 

 

 

 

 

 

@David Dotson @Trevor Gokey

General “reproducible computation” records and data infrastructure

High

 

 

Interoperable molecule class

 

 

 

@Simon Boothroyd @Joshua Horton

Bayesian infrastructure: ML frameworks

 

 

Bayesian Fitting

Analytically Differentiable System Object

 

 

BLOCKED

 

Off-site charges (support for conversion to other packages)

MEDIUM

 

 

Hard to spec without VirtualSite Handler implementation

 

 

 

 

Bespoke Torsion Parametrization package

 

 

 

Won’t be open source until fragmenter is refactored to be OE-free

Chemper conda-forge mir

CMILES conda-forge migration

QC submission infrastructure

 

 

IN PROGRESS PROTOTYPE

@Joshua Horton

Python System object (repo)

High

 

Analytical parameter gradient-based fitting, possibly other ML or Bayesian optimization routines

(Optional) spec from MolSSI interoperable molecule workgroup

OpenFF-core refactor

March 2020

 

IN PROGRESS PROTOTYPE

@Matt Thompson

CLI tool infrastructure

HIGH

 

 

 

June 2020

 

IN PROGRESS PROTOTYPE

@Matt Thompson @Jeffrey Wagner

Remove smirnoff_hack.py / Implement CachingToolkitWrappers

MEDIUM

 

 

 

 

 

 

 

Refactor Fragmenter / remove OE dependence / Base off OpenFF Molecule

High

 

 

Graph-based charges/WBOs

 

 

 

@David Dotson @John Chodera @Matt Thompson

OpenMM conda-forge migration

High

 

 

 

Late 2019

 

IN PROGRESS

@Jaime Rodríguez-Guerra Peter Eastman; Anthony Scopatz (contracted)l @John Chodera @Jeffrey Wagner Levi Naden

OpenFF Toolkit and general org repo conda-forge migration

High

 

 

OpenMM conda-forge migration

 

 

BLOCKED

 

Toolkit

Virtual Site ParameterHandler

MEDIUM

 

Off-site charge fitting

Likely to be reworked in the long-term to better work with the System object

March 2020

 

IN PROGRESS

@Trevor Gokey

WBO Bond ParameterHandler

HIGH

 

 

 

 

 

 

@Matt Thompson @David Dotson

AMBER-derived SMIRNOFF-format FF

HIGH

 

Biopolymer fitting

 

March 2020

 

IN PROGRESS

@David Cerutti @Jeffrey Wagner

Polarizability ParameterHandler

LOW

 

Polarizable fitting

 

 

 

 

 

A deep dive into toolkit parametrization differences (Josh Fass SMIRKS differences) / Automate complaining about cases where incoming molecule/chemistry is bad/misformatted

High

 

 

 

 

 

 

Spinoff (Potentially Shirts lab undergrad?)

Refactor/make our own Exception hierarchy, implement some problems as catch-able warnings.

MEDIUM

 

 

 

 

 

IN PROGRESS

@Matt Thompson @Simon Boothroyd @Jeffrey Wagner

Implement friendly default behavior, with option for custom validation logic when loading large datasets/high-volume pipelines. Consider making moleculefixer for common data problems.

MEDIUM

 

 

 

 

 

 

 

openforcefield-core/pydantic refactor (possibly driving a SMIRNOFF spec update)

High

 

 

Aromaticity refactor

Stereochemistry refactor

 

 

 

 

Protonation state enumeration

LOW

 

 

RDKit doesn’t have helpful protonation state enumeration; need to publicize and see if community wants to contribute there https://github.com/openforcefield/openforcefield/issues/526

Could use EPIK from schrodinger suite? Example in OpenMolTools

Mar 2020

July 2020 (incomplete)

PROTOTYPEBLOCKED

@Joshua Horton building on work of @Chaya Stern (Unlicensed)

RDKit stereochemistry and tautomer enumeration

High

 

 

 

 

 

 

@Jeffrey Wagner

Interoperable molecule/stereochemistry/aromaticity refactor

MEDIUM

 

 

Need to decide on desired behavior for how stereochemistry and aromaticity is handled. Also need to decide on which molecule formats should be losslessly round-trippable.

 

 

 

@Jeffrey Wagner

Biopolymer infrastructure (SMARTS typing optimization)

High , but can be after protein FF port

 

Biopolymer fitting

 

 

Dec 31 2020

 

@Jeffrey Wagner

Biopolymer infrastructure (infra improvement/Topology refactor/automated polymer unit recognition)

High , but can be after protein FF port

 

Biopolymer fitting

Should discuss design with OpenEye

 

 

 

 

Biopolymer infrastructure (graph charges and/or other scalable solution)

High , but can be after protein FF port

 

Biopolymer charge fitting

 

 

 

 

@Yuanqing Wang (Deactivated) @Josh Fass

CMAP torsions in OFFTK/SMIRNOFF spec

LOW

 

CMAP fitting

 

 

 

 

 

Fitting

Migrate FF optimization to ML framework

HIGH

 

 

 

 

 

 

 

Automate fitting infrastructure, remove OE dependencies

HIGH

 

 

QC Submission infrastructure (for QCMol-->graph mol conversion)

 

 

 

@Jeffrey Wagner @Hyesu Jang

PE parallelization (Fractalization? Key-value store in cloud? F@H? etc)

HIGH

 

 

MolSSI packaging Fractal separately from QCFractal

 

 

 

@David Dotson @John Chodera @Simon Boothroyd

Benchmarking

H-G benchmarking

Medium

 

 

 

 

 

 

 

P-L benchmarking (repo)

Medium

 

 

 

Mid 2019?

 

IN PROGRESS

@David Hahn

Property estimator mixed FF tests (mix AMBER and SMIRNOFF system components)

LOW

 

This may not be necessary.

 

 

 

 

 

Automated benchmarking + dashboard

May include geometry tools (MM minimization, conformer generation, torsion scanning, conformer scoring)

HIGH

 

 

(Optional) Reliable QCMol → OFFMol conversion/CMILES deviation checks

???

 

IN PROGRESS

Dashboard: @Jaime Rodríguez-Guerra @David Dotson @John Chodera @Trevor Gokey

Documentation / Community / Training

Docs cleanup and Binder-izing all examples

Medium

 

 

 

 

 

IN PROGRESS

@Matt Thompson (binder)

Developers guide + true community contributions/branch OE license issue resolution

HIGH

 

 

 

 

Dev docs will be a living document

IN PROGRESS

 

Training a 50% QCA developer

Medium

 

 

 

April 2020

 

IN PROGRESS

@David Dotson

Compute hosting for bespoke workflow on Hypernet Labs' Galileo Platform

Medium

 

 

Bespoke workflow prototype

 

 

IN PROGRESS

@Joshua Horton @Jeffrey Wagner

CHARMM-GUI integration

 

 

 

A way to create CHARMM residue template files (ParmEd Issue #1103)

 

 

BLOCKED

 

Pele-OpenFF Integration (meeting notes)

External

 

 

 

 

 

IN PROGRESS

Marti Municoy; Victor Guallar; @Jeffrey Wagner @David Mobley