How to override default charge assignment?

I am currently attempting to parameterize the PF6- anion with the Sage force field:

from openff.toolkit.typing.engines.smirnoff import ForceField
from openff.toolkit.topology import Molecule, Topology

pf6_minus = "F[P-](F)(F)(F)(F)F"

forcefield = ForceField("openff_unconstrained-2.0.0.offxml")
ethanol = Molecule.from_smiles(pf6_minus)
topology = Topology.from_molecules(molecules=[ethanol])
system = forcefield.create_openmm_system(topology)

but I am getting an UnassignedMoleculeChargeException.

Warning: Unable to load toolkit 'OpenEye Toolkit'. The Open Force Field Toolkit does not require the OpenEye Toolkits, and can use RDKit/AmberTools instead. However, if you have a valid license for the OpenEye Toolkits, consider installing them for faster performance and additional file format support: https://docs.eyesopen.com/toolkits/python/quickstart-python/linuxosx.html OpenEye offers free Toolkit licenses for academics: https://www.eyesopen.com/academic-licensing
[13:59:24] UFFTYPER: Warning: hybridization set to SP3 for atom 1
[13:59:24] UFFTYPER: Unrecognized charge state for atom: 1
/Users/orioncohen/miniconda3/envs/openmm/lib/python3.9/site-packages/openff/toolkit/typing/engines/smirnoff/parameters.py:4155: Warning: No registered toolkits can provide the capability "assign_partial_charges" for args "()" and kwargs "{'molecule': Molecule with name '' and SMILES '[F][P-]([F])([F])([F])([F])[F]', 'partial_charge_method': 'am1bcc', 'use_conformers': None, 'strict_n_conformers': False, 'normalize_partial_charges': True, '_cls': <class 'openff.toolkit.topology.molecule.FrozenMolecule'>}"
Available toolkits are: [ToolkitWrapper around The RDKit version 2021.03.5, ToolkitWrapper around AmberTools version 21.0, ToolkitWrapper around Built-in Toolkit version None]
 ToolkitWrapper around The RDKit version 2021.03.5 <class 'openff.toolkit.utils.exceptions.ChargeMethodUnavailableError'> : partial_charge_method 'am1bcc' is not available from RDKitToolkitWrapper. Available charge methods are ['mmff94'] 
 ToolkitWrapper around AmberTools version 21.0 <class 'TypeError'> : 'NoneType' object is not iterable
 ToolkitWrapper around Built-in Toolkit version None <class 'openff.toolkit.utils.exceptions.ChargeMethodUnavailableError'> : Partial charge method "am1bcc"" is not supported by the Built-in toolkit. Available charge methods are ['zeros', 'formal_charge']

  warnings.warn(str(e), Warning)
Traceback (most recent call last):
  File "/Users/orioncohen/projects/lowT/openmm/setup/testing_pf6.py", line 9, in <module>
    system = forcefield.create_openmm_system(topology)
  File "/Users/orioncohen/miniconda3/envs/openmm/lib/python3.9/site-packages/openff/toolkit/typing/engines/smirnoff/forcefield.py", line 1348, in create_openmm_system
    parameter_handler.postprocess_system(system, topology, **kwargs)
  File "/Users/orioncohen/miniconda3/envs/openmm/lib/python3.9/site-packages/openff/toolkit/typing/engines/smirnoff/parameters.py", line 3944, in postprocess_system
    raise UnassignedMoleculeChargeException(msg)
openff.toolkit.typing.engines.smirnoff.parameters.UnassignedMoleculeChargeException: The following molecules did not have charges assigned by any ParameterHandler in the ForceField:
[F][P-]([F])([F])([F])([F])[F]

I have RESP fit charges for PF6-. Can I make ForceField use those parameters so it doesn’t throw an error?

[orionarcher]

Update, I believe that I found the error. In my attempt to support arbitrary permutations of atom ordering, I made some accidental arbitrary permutations.

Iterating through Forces does follow the same Atom ordering as Topology and System.

Thank you for the help everyone and apologies for the mistake!

[mattwthompson]

The quick-and-dirty approach is to assign your charges to the Moleucle object via its partial_charges setter and then instruct ForceField.create_openmm_system to use those parameters. Be aware this might be deprecated in the future since it muddies the water of what a force field is intended to do (#806).

import numpy as np
from openmm import NonbondedForce, unit

from openff.toolkit.topology import Molecule
from openff.toolkit.typing.engines.smirnoff import ForceField

pf6_minus = Molecule.from_smiles("F[P-](F)(F)(F)(F)F")

# Assign partial charges from a list-like object - replace with numbers from your favorite method
pf6_minus.partial_charges = np.asarray([0, -1, 0, 0, 0, 0, 0]) * unit.elementary_charge

forcefield = ForceField("openff_unconstrained-2.0.0.offxml")
topology = pf6_minus.to_topology()

system = forcefield.create_openmm_system(topology, charge_from_molecules=[pf6_minus])

# Inspect the openmm.System object to see which charges 
# actually ended up on each particle
for force in system.getForces():
    if type(force) == NonbondedForce:
        for i in range(force.getNumParticles()):
            print(force.getParticleParameters(i)[0])

A more programmatic approach is to modify the force field itself to include your pre-specified charges. This method (LibraryChargeType.from_molecule) creates a parameter whose SMIRKS pattern is unique to your molecule and also contains those charges.

import numpy as np
from openmm import NonbondedForce, unit

from openff.toolkit.topology import Molecule
from openff.toolkit.typing.engines.smirnoff import ForceField
from openff.toolkit.typing.engines.smirnoff.parameters import LibraryChargeHandler

pf6_minus = Molecule.from_smiles("F[P-](F)(F)(F)(F)F")

# Assign partial charges from a list-like object
pf6_minus.partial_charges = np.asarray([0, -1, 0, 0, 0, 0, 0]) * unit.elementary_charge

library_charge_type = LibraryChargeHandler.LibraryChargeType.from_molecule(pf6_minus)

forcefield = ForceField("openff_unconstrained-2.0.0.offxml")
forcefield["LibraryCharges"].add_parameter(parameter=library_charge_type)

topology = pf6_minus.to_topology()

system = forcefield.create_openmm_system(topology)

# Inspect the openmm.System object to see which charges
# actually ended up on each particle
for force in system.getForces():
    if type(force) == NonbondedForce:
        for i in range(force.getNumParticles()):
            print(force.getParticleParameters(i)[0])

Either approach should produce the same result:

$ python library_charges.py
0.0 e
-1.0 e
0.0 e
0.0 e
0.0 e
0.0 e
0.0 e

Interchange should be able to process these objects as well:

# copy everything up to `system = forcefield.create_openmm_system`
...
from openff.interchange.components.interchange import Interchange

out = Interchange.from_smirnoff(force_field=forcefield, topology=topology)
out.to_whatever("file.txt")  # GROMACS, Amber, LAMMPS, etc.

but there may be some bugs; please carefully validate anything that Interchange exports if you use it in research work.