Protein Force Field Project Plan

Objective

Derive force field parameters for proteins consistent with the OpenFF small molecule force field.

Primary Driver

@Chapin Cavender

Approvers

@Michael Gilson @Michael Shirts

Supporting Drivers

 

Stakeholders (these people will be tagged in project update notifications on Slack)

@Jeffrey Wagner @Pavan Behara @Iván Pulido @Lily Wang @Joshua Horton @David Mobley

Project Manager

@Diego Nolasco

Page Owner (only this person can edit this page)

@Chapin Cavender

Decision authority

Majority of Primary Driver and all Approvers, only in “Biopolymer FF call” meetings

Veto authority: Primary Driver, any Approver

Discussion/notification venue

Fortnightly “Biopolymer FF call” meetings (decision forum). NOT “FF release call” meetings.

#ff-biopolymers channel on OpenFF slack and “FF release call” meetings (notification and discussion, no major decisions allowed here. It is not assumed that meeting attendees have read the slack discussions, they must be summarized during meetings to be considered in decisions)

Meeting notes

https://openforcefield.atlassian.net/wiki/spaces/MEET/pages/2250932319

Due date

2022-01-01

Key outcomes

  • Extensible SMIRNOFF format for amino acid residues

  • Training datasets for 20 natural amino acids

  • Selection of FF model for proteins

  • One or more sets of OpenFF parameters for the 20 canonical amino acids

  • Identification of key benchmark systems

Status

PLANNING PHASE

SMIRNOFF format

We need SMARTS strings that can specify protein-specific terms for general amino acids. To summarize the discussion here: https://openforcefield.atlassian.net/wiki/pages/createpage.action?spaceKey=MEET&title=2020-04-01%20AMBER%20FF%20porting%20meeting%20notes

  • Amber ff14SB was ported to SMIRNOFF format by using SMARTS strings that capture an entire amino acid, differentiating between main chain and terminal residues and between protonation/tautomeric states

  • Previous approach is not extensible for modified or synthetic amino acids

  • Need general SMARTS strings for backbone and side chain torsions in polypeptide chains

  • [#6X3](=O)-[#7X3:1]-[#6X4:2]-[#6X3:3](=O)-[#7X3H1:4]-[#6X4] will tag ψ for all residues except proline

Training datasets

This is a list of datasets that could be used to train protein-specific force field parameters.

Relevant small molecule datasets from Parsley

Needs to be determined. In particular, did Parsley train on dipeptides or tripeptides for any of the 20 canonical amino acids?

Dataset name

Dataset type

QC method

Molecules

QCA submission

Dataset name

Dataset type

QC method

Molecules

QCA submission

Dipeptides

 

 

 

 

Tripeptides

 

 

 

 

Protein-specific datasets

Cerutti tetrapeptides are a set of 185 tetrapeptides with sidechains X-Y-X, X in [Ala, Gly, Ser, Val], and Y in [Ala, Arg, Ash, Asn, Asp, Cys, Glh, Gln, Glu, Gly, Hid, Hie] excluding (X == Ser && Y == Glu). David Cerutti selected multiple conformers for each tetrapeptide.

Dataset name

Dataset type

QC method

Molecules

QCA submission

Status

Dataset name

Dataset type

QC method

Molecules

QCA submission

Status

OpenFF Protein Fragments Initial

Optimization

B3LYP-D3BJ/def2-TVPP

16 tetrapeptides with sidechains X-Ala-X and X in [Ala, Gly, Ser, Val]

2020-07-06-OpenFF-Protein-Fragments-Initial

Complete

OpenFF Protein Fragments version 2

Optimization

B3LYP-D3BJ/DZVP

Cerutti tetrapeptides with constraints to avoid hydrogen bonds

2020-08-12-OpenFF-Protein-Fragments-version2

Superseded

OpenFF Protein Peptide Fragments constrained v1.0

Optimization

B3LYP-D3BJ/DZVP

Cerutti tetrapeptides with constraints to avoid hydrogen bonds

2020-08-12-OpenFF-Protein-Fragments-version2

Complete

OpenFF Protein Peptide Fragments unconstrained v1.0

Optimization

B3LYP-D3BJ/DZVP

Cerutti tetrapeptides with no constraints

2020-10-27-OpenFF-Protein-Fragments-Unconstrained

ERRORED

OpenFF Protein Fragments TorsionDrives v1.0

TorsionDrives on ϕ, ψ, ω, χ1, and χ2

B3LYP-D3BJ/DZVP

Cerutti tetrapeptides

2020-09-16-OpenFF-Protein-Fragments-TorsionDrives

ERRORED

(22 / 845 errored)

OpenFF PEPCONF OptimizationDataset v1.0

Optimization

B3LYP-D3BJ/DZVP

PEPCONF dataset

2020-10-26-PEPCONF-Optimization

ERRORED

(6000 / 7560 errored)

OpenFF Benchmark Ligands

TorsionDrive

B3LYP-D3BJ/DZVP

OpenFF FEP benchmark

2020-07-27-OpenFF-Benchmark-Ligands

Complete

Model

We envision several tiers of models, presented below in order of increasing anticipated effort. We will generate and benchmark lower-effort models first and use the results of the benchmarks to inform decisions about higher-effort models.

Null Model

The null model is that the small molecule force field already describes proteins well and needs no protein-specific parameters. Valence parameters (bonds and angles), torsions, and Lennard-Jones parameters will be copied from the small molecule force field. Parsley was trained on compounds that resemble protein backbone and sidechain analogs, so these parameters are likely a good first pass at describing polypeptide chains. Partial charges could be derived in several ways:

  • Copy library charges from existing protein force field: Amber ff14SB (RESP, unchanged from Amber ff99) or Amber ff15ipq (IPolQ)

    • Pro: easy to implement

    • Pro: we know these work pretty well in the Amber context

    • Con: we are no longer in the Amber context

  • Generate library charges for the 20 canonical amino acids (main chain and terminal) using AM1-BCC (RESP2)

    • Pro: consistent with Parsley; for example, we want the parameters of a serine side chain to look a lot like those of ethanol, since we have reason to believe these parameters play well with the other parameters in the FF

    • Pro: @Lily Wang has evidence that AM1-BCC charges of fragments are similar (< 0.1 e) to the charges from a larger polymer (see https://zenodo.org/record/4977401#.YNuCk34pCpp )

    • Con: more effort to generate

  • Generate charges on-the-fly using graph convolutional networks (see https://openforcefield.atlassian.net/wiki/pages/createpage.action?spaceKey=MEET&title=2020-04-01%20AMBER%20FF%20porting%20meeting%20notes)

    • Maybe don’t take this approach unless/until it is also being used for the small molecules

Protein-specific Torsion (PST) Model

The Protein-specific Torsion Model includes protein-specific torsion terms, while other valence parameters (bonds and angles) and Lennard-Jones parameters will be copied from the small molecule force field. Charges will be derived using one of the methods described in the Null Model. Protein-specific torsions for the backbone torsions (ϕ, ψ, and ω) and sidechain torsions (χ1, maybe χ2) will be derived by fitting to QC datasets using dipeptides, tripeptides, and tetrapeptides. The dihedral angles with protein specific terms will be a subset of:

  • Flexible backbone dihedrals (ϕ and ψ)

  • Peptide bond dihedral (ω)

  • First sidechain dihedral (χ1)

  • Second sidechain dihedral (χ2)

A major decision for this model is which sidechains should have unique torsions that override the general peptide backbone torsion. We envision using the Protein Fragments Optimization and TorsionDrive datasets as the primary training data. Then, other datasets such as PEPCONF can be used as validation data to make decisions about the model. Alternatively, automated chemical perception (https://openforcefield.atlassian.net/wiki/spaces/FF/pages/10387503) may be used to identify dihedrals that are not described well in the Parsley training set. The resulting model will likely be the candidate for the first protein force field release.

CMAP model

The CMAP model includes a protein-specific correction map (CMAP) that fits a 2D potential in ϕ and ψ to QC datasets. This model uses the same valence parameters, Lennard-Jones parameters, and charges as the Torsion Model. The largest obstacle to generating this model is specifying the CMAP potential in the SMIRNOFF format.

Protein-specific Torsions & Lennard-Jones (PSTLJ) Model

The Protein-specific Torsions & Lennard-Jones Model includes protein-specific terms for torsions (as in the PST Model) and protein-specific Lennard-Jones parameters. Valence parameters (bonds and angles) will be copied from the small molecule force field. Charges will be derived using one of the methods described in the Null Model. Protein-specific torsions will be derived in the same way as the PST Model or the CMAP model. Protein-specific Lennard-Jones parameters will be derived by fitting to experimental data for dipeptides or small molecule analogs. It is likely that some curation will be necessary to learn what data is available, but examples include:

A major decision for this model is which set of Lennard-Jones types should be fit. This choice will be helped immensely by automated Lennard-Jones typing using Monte Carlo sampling, which is still under development.

Benchmarks

Experimental datasets are being curated to evaluate protein force fields. These datasets will be published as a LiveCoMS review, described here: https://openforcefield.atlassian.net/wiki/spaces/COMMS/pages/1927413777. It will be useful to identify a small number of key benchmarks that can interrogate distinct physical properties of proteins and that can be completed relatively quickly (~1 month). These key benchmarks will be used to evaluate force field models and make decisions about more complex models.

Milestones and deadlines

Milestone

Owner

Deadline

Status

Notes

Milestone

Owner

Deadline

Status

Notes

Generate Null Model with Amber library charges

@Chapin Cavender

2021-07-09

NOT STARTED

 

Generate Null Model with AM1-BCC library charges

@Chapin Cavender

2021-08-01

NOT STARTED

Waiting on infrastructure for getting polymer charges from fragments

Choose key benchmarks to quickly evaluate force field models

Biopolymer FF group

2021-08-01

IN PROGRESS

In parallel with LiveCoMS review

Run key benchmarks for Null Models with library charges

@Chapin Cavender

Maybe others

2021-10-01

NOT STARTED

 

Decide on QC data for PST Model

Biopolymer FF group

2021-08-01

IN PROGRESS

Started by Dave Cerutti

Run QC calculations for PST Model

@Chapin Cavender

2021-10-01

IN PROGRESS

Started by @David Dotson and @Trevor Gokey

Fit PST Model with one general term for all sidechains

@Chapin Cavender

2021-11-01

NOT STARTED

 

Decide on sidechain-specific terms for PST Model

Biopolymer FF group

2021-11-01

NOT STARTED

 

Fit PST Model with sidechain-specific terms

@Chapin Cavender

2022-01-01

NOT STARTED

 

Run key benchmarks for PST Models

@Chapin Cavender

Maybe others

2022-03-01

NOT STARTED

 

Generate charges using graph convolutional networks

@Chapin Cavender

2022

NOT STARTED

Need update on feasibility from Chodera group

Fit CMAP model, if necessary

@Chapin Cavender

2022

NOT STARTED

Waiting on CMAP infrastructure

Fit PSTLJ model, if necessary

@Chapin Cavender

2022

NOT STARTED

Manual or automated LJ typing