Skip to content

cli_tutorial.md

Latest commit

 

History

History
349 lines (267 loc) · 16.2 KB

cli_tutorial.md

File metadata and controls

349 lines (267 loc) · 16.2 KB

Binding Free Energies with the OpenFE CLI

This tutorial demonstrates how to use the OpenFE CLI (Command Line Interface) to calculate free energies - with no Python at all!

The CLI is useful for simple setups, but you may need to use the Python API for more complicated setups.

RBFE calculations with openfe are split into 3 steps: plan, run, and gather, each of which corresponds to a CLI command:

  1. openfe plan-rbfe-network: Define the systems and prepare simulations to be run.
  2. openfe quickrun: Run the simulations
  3. openfe gather: Gather and analyze simulation results to generate a table of free energies.

0. Collect input files

To work through this tutorial, start out with a fresh directory.

You can download the tutorial materials (including these instructions) using the command:

openfe fetch rbfe-tutorial

Then when you run ls, you should see that your directory has:

  • cli_tutorial.md: the file containing these instructions
  • python_tutorial.ipynb: a notebook detailing how to do this analysis using the Python API, instead of the CLI shown here.
  • tyk2_ligands.sdf and tyk2_protein.pdb : files containing the molecules we'll use in this tutorial.

1. Set up the campaign

The CLI makes setting up the simulation very easy - it's just a single CLI command. There are separate commands for relative binding free energy (RBFE) and relative hydration free energy setups (RHFE).

For RBFE campaigns, the relevant command is openfe plan-rbfe-network. For RHFE, the command is openfe plan-rhfe-network. They work mostly the same, except that the RHFE planner does not take a protein. In this tutorial, we'll perform an RBFE calculation. The only difference for RHFE is in the setup stage - running the simulations and gathering the results are the same.

The single command:

openfe plan-rbfe-network -M tyk2_ligands.sdf -p tyk2_protein.pdb -o network_setup/ --n-protocol-repeats 1

performs the following steps:

  • Read all the ligands from the SDF by giving the option -M tyk2_ligands.sdf. You can also use -M with a directory, and it will load all molecules found in any SDF or MOL2 file in that directory.

  • Pass a PDB of the protein target (TYK2) with -p tyk2_protein.pdb.

  • Create transformation JSONs, stored in the directory network_setup/, that contain all information needed to run simulations with openfe quickrun.

  • Instruct openfe to only run one repeat of the alchemical simulation per quickrun call using --n-protocol-repeats 1.

    Note: openfe's default behaviour is to use three repeats to calculate the uncertainty (i.e. standard deviation) in an estimate. When setting --n-protocol-repeats 1, you must execute the transformation multiple times - at minimum 2, but best practice is 3 independent repeats.

Planning the campaign may take some time due to the complex series of tasks involved:

  • partial charges are generated for each of the ligands to ensure reproducibility, by default this requires a semi-empirical quantum chemical calculation to calculate am1bcc charges
  • atom mappings are created and scored based on the perceived difficulty for all possible ligand pairs
  • an optimal network is extracted from all possible pairwise transformations which balances edge redundancy and the total difficulty score of the network

The partial charge generation can take advantage of multiprocessing which offers a significant speed-up, you can specify the number of processors available using the -n flag:

openfe plan-rbfe-network -M tyk2_ligands.sdf -p tyk2_protein.pdb -o network_setup --n-protocol-repeats 1 -n 4

This will result in a directory called network_setup/, which is structured like this:

network_setup
├── ligand_network.graphml
├── network_setup.json
└── transformations/
    ├── rbfe_lig_ejm_31_complex_lig_ejm_42_complex.json
    ├── rbfe_lig_ejm_31_complex_lig_ejm_46_complex.json
    ├── rbfe_lig_ejm_31_complex_lig_ejm_47_complex.json
    ├── rbfe_lig_ejm_31_complex_lig_ejm_48_complex.json
    ├── rbfe_lig_ejm_31_complex_lig_ejm_50_complex.json
    ├── rbfe_lig_ejm_31_solvent_lig_ejm_42_solvent.json
    ├── rbfe_lig_ejm_31_solvent_lig_ejm_46_solvent.json
    ...

The ligand_network.graphml file describes the network of ligands connected by atom mappings.

We can visualize this network with the openfe view-ligand-network command to open an interactive viewer.:

openfe view-ligand-network network_setup/ligand_network.graphml

You can move the ligand names around to get a better view of the structure, and if you click on the edge, you will see the mapping for that edge.

The files that describe each individual simulation we will run are located within network_setup/transformations/. Each JSON file represents a single alchemical leg to run and contains all the necessary information to run that leg. Filenames indicate ligand names as taken from the SDF; for example, the file rbfe_lig_ejm_31_complex_lig_ejm_42_complex.json is the leg associated with the transformation of the ligand lig_ejm_31 into lig_ejm_42 while in complex with the protein.

A single RBFE between a pair of ligands requires running two legs of an alchemical cycle (JSON files) - one for the ligand in solvent, and one for the ligand complexed with the protein. The results from these two simulations can then be combined in the next step (openfe gather) to obtain a single $\Delta\Delta G$ relative binding free energy value.

Note that this specific setup makes a number of choices for you, from filenames to default values. All of these choices can be customized in the Python API. Here are the specifics on how these simulation are set up:

  1. kartograf is used to generate the atom mappings between ligands.
  2. The ligand network is a minimal spanning tree, with the default LOMAP scorer used to score the mappings.
  3. Solvent is water with NaCl at an ionic strength of 0.15 M (neutralized) with a minimum distance of 1.2 nm from the solute to the edge of the box.
  4. The protocol used is OpenFE's OpenMM-based Hybrid Topology RFE protocol, with default settings.

Optional step: Customize your campaign setup

OpenFE contains many different options and methods for setting up a simulation campaign. While less flexible than using the API, some options can be modified by providing a settings file in the .yaml format.

The default settings represented in YAML settings format is as follows:

mapper: kartograf
    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
        allow_bond_breaks: false

network:
    method: generate_minimal_spanning_network

partial_charge:
    method: am1bcc
    settings:
        off_toolkit_backend: ambertools
        number_of_conformers: None
        nagl_model: None

Let's assume you want to exchange the kartograf atom mapper with the LOMAP atom mapper, the Minimal Spanning Tree Network Planner with the Maximal Network Planner and the am1bcc charge method with OpenFF NAGL:

Provide a file like settings.yaml with the desired changes:

mapper:
  method: lomap

network:
  method: generate_maximal_network

partial_charge:
  method: nagl
  settings:
    nagl_model: null  # null specifies the use of the latest nagl model

Plan your rbfe network with an additional -s flag for passing the settings:

openfe plan-rbfe-network -M tyk2_ligands.sdf -p tyk2_protein.pdb -o network_setup --n-protocol-repeats 1 -s settings.yaml

The output of the CLI program will now reflect the changes made:

RBFE-NETWORK PLANNER
______________________

Parsing in Files:
	Got input:
		Small Molecules: SmallMoleculeComponent(name=lig_ejm_31) SmallMoleculeComponent(name=lig_ejm_42) SmallMoleculeComponent(name=lig_ejm_43) SmallMoleculeComponent(name=lig_ejm_46) SmallMoleculeComponent(name=lig_ejm_47) SmallMoleculeComponent(name=lig_ejm_48) SmallMoleculeComponent(name=lig_ejm_50) SmallMoleculeComponent(name=lig_jmc_23) SmallMoleculeComponent(name=lig_jmc_27) SmallMoleculeComponent(name=lig_jmc_28)
		Protein: ProteinComponent(name=)
		Cofactors: []
		Solvent: SolventComponent(name=O, Na+, Cl-)

Using Options:
	Mapper: <LomapAtomMapper (time=20, threed=True, max3d=1.0, element_change=True, seed='', shift=False)>
	Mapping Scorer: <function default_lomap_score at 0x166bc5300>
	Network Generation: <function generate_minimal_spanning_network at 0x16a413e20>
	Partial Charge Generation: am1bcc

	n_protocol_repeats=1 (1 simulation repeat(s) per transformation)

To see all settings customizable by YAML input, run openfe plan-rbfe-network -h.

2. Run the simulations

For this tutorial, we have precalculated data you can load, since running the simulations can take a long time. However, you could, in principle, run each simulation on your local machine.

You can run each leg individually by using the openfe quickrun command:

openfe quickrun path/to/transformation.json -o results.json -d working-directory

where

  • path/to/transformation.json is the path to one of the transformation files created by openfe plan-rbfe-network in the prior step.
  • -o results.json to give the final output JSON file and -d for the directory where simulation results should be stored.

To run one simulation from the tutorial data, a command might look like:

openfe quickrun network_setup/transformations/rbfe_lig_ejm_31_solvent_lig_ejm_42_solvent.json -o results/rbfe_lig_ejm_31_solvent_lig_ejm_42_solvent.json -d results/rbfe_lig_ejm_31_solvent_lig_ejm_42_solvent/

When running a complete network of simulations, it is important to ensure that the file name for the result JSON and name of the working directory are different for each leg and each repeat, otherwise you'll overwrite results. We recommend doing this programmatically, such as the example below, which uses the fact that the JSON files in network_setup/transformations/ have unique names, and creates directories and result JSON files based on those names.

To run all legs sequentially (not recommended!!) you could do something like:

# this will take a very long time! don't actually do it!
for file in network_setup/transformations/*.json; do
  relpath=${file:30}  # strip off "network_setup/transformations/"
  dirpath=${relpath%.*}  # strip off final ".json"
  # loop over three repeats
  for repeat in {1..3}; do
      openfe quickrun $file -o results/repeat${repeat}/$relpath -d results/repeat${repeat}/$dirpath
  done
done

In practice, you probably want to submit the simulations to an HPC queue. In that case, you'll want to create a new job script for each simulation JSON file, and the core of that job script will be to run the openfe quickrun command above.

Details of what information is needed in that job script will depend on your computing center, but below is an example of a very simple script that will create and submit a job script for the simplest SLURM use case:

for file in network_setup/transformations/*.json; do
  relpath=${file:30}  # strip off "network_setup/transformations/"
  dirpath=${relpath%.*}  # strip off final ".json"
  for repeat in {1..3}; do
      jobpath="network_setup/transformations/${dirpath}_${repeat}.job"
      cmd="openfe quickrun $file -o results/repeat${repeat}/$relpath -d results/repeat${repeat}/$dirpath"
      echo -e "#!/usr/bin/env bash\n${cmd}" > $jobpath
      sbatch $jobpath
  done
done

The approach listed here is what was used for the example results that we'll download in the next section.

3. Gather the results

To get example simulation output data, use the following commands:

openfe fetch rbfe-tutorial-results
tar xzf rbfe_results.tar.gz

This will create a directory called results/ that contains files with the file structure you would get from running the calculations as above. The result JSON files are the actual results of a simulation. To keep this example data a reasonable size, files typically generated during the simulation (such as detailed simulation information) have been replaced by empty files to keep the size smaller. The structure should look something like this:

results
├── replicate_0
│   ├── rbfe_lig_ejm_31_complex_lig_ejm_42_complex
│   │   ├── shared_RelativeHybridTopologyProtocolUnit-79c279f04ec84218b7935bc0447539a9_attempt_0
│   │   │   ├── checkpoint.nc
│   │   │   ├── db.json
│   │   │   ├── simulation_real_time_analysis.yaml
│   │   │   └── simulation.nc
│   │   ├── shared_RelativeHybridTopologyProtocolUnit-a3cef34132aa4e9cbb824fcbcd043b0e_attempt_0
│   │   │   ├── checkpoint.nc
│   │   │   ├── db.json
│   │   │   ├── simulation_real_time_analysis.yaml
│   │   │   └── simulation.nc
│   │   └── shared_RelativeHybridTopologyProtocolUnit-abb2b104151c45fc8b0993fa0a7ee0af_attempt_0
│   │       ├── checkpoint.nc
│   │       ├── db.json
│   │       ├── simulation_real_time_analysis.yaml
│   │       └── simulation.nc
│   ├── rbfe_lig_ejm_31_complex_lig_ejm_42_complex.json
│   ├── rbfe_lig_ejm_31_complex_lig_ejm_46_complex
│   │   ├── shared_RelativeHybridTopologyProtocolUnit-361500fe831c431aa830efd207db0955_attempt_0
│   │   │   ├── checkpoint.nc
│   │   │   ├── db.json
│   │   │   ├── simulation_real_time_analysis.yaml
│   │   │   └── simulation.nc
│   │   ├── shared_RelativeHybridTopologyProtocolUnit-5a6176cfbf074f92bc76caac91b1c1bf_attempt_0
│   │   │   ├── checkpoint.nc
│   │   │   ├── db.json
│   │   │   ├── simulation_real_time_analysis.yaml
│   │   │   └── simulation.nc
│   │   └── shared_RelativeHybridTopologyProtocolUnit-e16de73f07964e9096f34611e0c874ca_attempt_0
│   │       ├── checkpoint.nc
│   │       ├── db.json
│   │       ├── simulation_real_time_analysis.yaml
│   │       └── simulation.nc
│   ├── rbfe_lig_ejm_31_complex_lig_ejm_46_complex.json
...

The JSON results file contains not only the calculated $\Delta G$, and uncertainty estimate, but also important metadata about what happened during the simulation. In particular, it will contain information about any errors or failures that occurred -- these errors will not cause the entire campaign to fail, and will be recorded so you can later analyze what went wrong.

To gather all the $\Delta G$ estimates into a single file, use the openfe gather command from within the working directory used above:

openfe gather results/ --report dg -o final_results.tsv

Note that if you have multiple results directories, you can pass multiple directories, e.g. openfe gather results_0/ results_1/.

This will write out a tab-separated table of results where the results reported are controlled by the --report option:

  • dg (default) reports the ligand and the results are the maximum likelihood estimate of its absolute free, and the associated uncertainty from DDG replica averages and standard deviations.
  • ddg reports pairs of ligand_i and ligand_j, the calculated relative free energy DDG(i->j) = DG(j) - DG(i) and its uncertainty.
  • raw reports the raw results, giving the leg (vacuum, solvent, or complex), ligand_i, ligand_j, the raw DG(i->j) associated with it.

The resulting file (final_results.tsv) will look something like this:

ligand	DG(MLE) (kcal/mol)	uncertainty (kcal/mol)
lig_ejm_31	-0.09	0.05
lig_ejm_42	0.7	0.1
lig_ejm_46	-0.98	0.05
lig_ejm_47	-0.1	0.1
lig_ejm_48	0.53	0.09
lig_ejm_50	0.91	0.06
lig_ejm_43	2.0	0.2
lig_jmc_23	-0.68	0.09
lig_jmc_27	-1.1	0.1
lig_jmc_28	-1.25	0.08