plan-rbfe-network command#

openfe plan-rbfe-network#

Plan a relative binding free energy network, saved as JSON files for use by the quickrun command.

This tool is an easy way to set up a RBFE calculation campaign. The JSON files this outputs can be used to run each leg of the campaign. openfe. The generated Network will be stored in a folder containing for each transformation a JSON file, that can be run with quickrun.

Note

To ensure a consistent set of partial charges are used for each molecule across different transformations, this tool will automatically generate charges ahead of planning the network. am1bcc charges will be generated via ambertools, this can also be customized using the settings yaml file.

By default, this tool makes the following choices:

  • Atom mappings performed by KartografAtomMapper, with settings atom_max_distance=0.95, atom_map_hydrogens=True, map_hydrogens_on_hydrogens_only=True, map_exact_ring_matches_only=True, allow_partial_fused_rings=True, and allow_bond_breaks=False.

  • Minimal spanning network as the network planner, with LOMAP default score as the weight function

  • Water as solvent, with NaCl counter ions at 0.15 M concentration.

  • Protocol is the OpenMM-based relative hybrid topology protocol, with default settings.

These choices can be customized by creating a settings yaml file, which is passed in via the -s settings.yaml option, which is detailed in the Options section. For more advanced setups, please consider using the Python layer of openfe.

openfe plan-rbfe-network [OPTIONS]

Options

-M, --molecules <molecules>#

Required A directory or file containing all molecules to be loaded, either as a single SDF or multiple MOL2/SDFs. Any number of sdf paths.

-p, --protein <protein>#

Required ProteinComponent. Can be provided as an PDB or as a PDBx/mmCIF file. string.

-C, --cofactors <cofactors>#

Path to cofactors sdf file. This may contain multiple molecules

-s, --settings <yaml_settings>#

Path to a YAML file specifying the atom mapper (mapper), network planning algorithm (network), and/or partial charge method (partial_charge) to use.

Supported atom mapper choices are:
- KartografAtomMapper (default as of v1.7.0)
- LomapAtomMapper
Supported network planning algorithms include (but are not limited to):
- generate_minimal_spanning_network (default)
- generate_minimal_redundant_network
- generate_radial_network
- generate_lomap_network
Supported partial charge method choices are:
- am1bcc (default)
- am1bccelf10 (only possible if off_toolkit_backend is openeye)
- nagl (must have openff-nagl installed)
- espaloma (must have espaloma_charge installed)

settings: allows for passing in any keyword arguments of the method’s corresponding Python API.

For example:

mapper:
  method: LomapAtomMapper
  settings:
    element_change: false

network:
  method: generate_minimal_redundant_network
  settings:
    mst_num: 3

partial_charge:
  method: am1bcc
  settings:
    off_toolkit_backend: ambertools
-o, --output-dir <output_dir>#

Path to the output directory. Defaults to ./alchemicalNetwork.

--n-protocol-repeats <n_protocol_repeats>#

Number of independent repeat(s) to be run per execution of a transformation using the openfe quickrun command.

For example:

--n-protocol-repeats=3 means openfe quickrun will execute 3 repeats in serial.

--n-protocol-repeats=1 means openfe quickrun will execute only 1 repeat per call, which allows for individual repeats to be submitted in parallel by calling openfe quickrun on the same input JSON file multiple times.

-n, --n-cores <n_cores>#

Number of cores to use for multiprocessing.

--overwrite-charges#

Overwrite any partial charges present in the input molecules.

On this page

This Page