Suggestions for new examples

[j-wags]

This Issue is intended to keep track of the examples that we could implement in the future. Please vote on these using 👍 if they would be useful to you.

[j-wags]

An example to compare parameters assigned to two separate systems, from a variety of sources (OpenMM systems, amber prmtops, gromacs, etc)

Rationale: Somebody from industry may want to translate their own parameters (say a frcmod file) into SMIRNOFF format. In this case, they may have leap-parametrized files, and comparing the two systems may be helpful to verify that they translated their parameters correctly.

It would be similar to this, but more transparent about where the non-OFF parameters can come from. It would also include examples of comparing parameters from formats other than AMBER.

[j-wags]

An example to show how to read, manipulate, and write cosmetic parameters.

For reference: https://github.com/openforcefield/openforcefield/issues/160#issuecomment-480047055

[j-wags]

Add a 2D depiction to examples/inspect_assigned_parameters

We had this functionality previously in https://github.com/openforcefield/openforcefield/blob/master/utilities/convert_frosst/parameter_usage.ipynb

It might also be a good first issue for someone to implement this in RDKit.

[j-wags]

~An example showing how to load charges from molecules, similar to this test~

Turned into its own issue at #251

[j-wags]

An example showing how to query the available cheminformatics toolkits and switch their priorities

[andrrizzi]

An example starting from AMBER files and swapping the parameters of one of its molecules with SMIRNOFF parameters.

[davidlmobley]

An example starting from AMBER files and swapping the parameters of one of its molecules with SMIRNOFF parameters.

This may not be possible (or at least, not something we want to have be possible) without also having mol2/sdf files of the component molecules: prmtop/crd files don't contain information on bond types, formal charge, etc. Reconstructing the molecules from them would be a pain, I think.

[j-wags]

An example that minimizes structures in SD files (+2 from Alberto Gobbi and Ian Craig)

Since we were just talking about workflows on today's advisory board call, I'd like to support the request for an example that minimizes structures in SD files (from Alberto Gobbi, Jan 23rd above).

[proteneer]

An example showing how to set up a simple free energy calculation (either relative or absolute) on either a host-guest system or more ideally, a protein-ligand system.

[slochower]

An example starting from AMBER files and swapping the parameters of one of its molecules with SMIRNOFF parameters.

I have this.

[slochower]

An example showing how to set up a simple free energy calculation (either relative or absolute) on either a host-guest system or more ideally, a protein-ligand system.

This also is out there in a few different places, depending on what you want. Do you want the method of setting up a free energy calculation calculation (in general) or just how to apply SMIRNOFF99Frosst parameters to host-guest systems (https://github.com/openforcefield/openforcefield/blob/master/examples/using_smirnoff_with_amber_protein_forcefield/BRD4_inhibitor_benchmark.ipynb and https://github.com/slochower/smirnoff-host-guest)?

[ppxasjsm]

Having gone through the available examples I feel like it is not massively easy to get started with openff simulations, particularly if you come from an antechamber tleap or similar world. It isn't really clear, how I would load a small molecule, parametrise it and then solvate it and then save it for later use.

[davidlmobley]

@ppxasjsm Those are good points, though if you're used to antechamber or tleap your best bet would really just be to keep doing things the way you normally do, completely set up your system, then swap in new parameters for your ligand/solute as per this example: https://github.com/openforcefield/openforcefield/blob/master/examples/swap_amber_parameters/swap_existing_ligand_parameters.ipynb

[jchodera]

Thanks for the feedback, @ppxasjsm! If you're willing to use the nightly openmm dev build from omnia-dev, you can use the small molecule residue template generators or unified SystemGenerator in openmmforcefields reassign SMIRNOFF parameters to small molecules while using AMBER biopolymer and solvent models easily. There is a simplified example in this PR (waiting until the new OpenMM version is released): https://github.com/openforcefield/openforcefield/pull/486