ABFE¶
This document describes how to run a ABFE simulation using Deep Origin tools.
Prerequisites¶
We assume that we have an initialized and configured Complex object:
from deeporigin.drug_discovery import Complex, EXAMPLE_DATA_DIR
sim = Complex.from_dir(EXAMPLE_DATA_DIR)
Here, ABFE requires that the Complex object have an already prepared protein (PDB), and the associated ligands (SDF) are in a docked pose.
Warning
The Complex.from_dir() function only accepts 1 PDB file per directory. This function will throw an error if it finds more than 1 PDB file per directory.
For more details on how to get started, see Getting Started .
Starting an ABFE run¶
Single ligand¶
To run an end-to-end ABFE workflow on a single ligand, we use:
jobs = sim.abfe.run_end_to_end(ligands=[sim.ligand[0]]) # for example
job = jobs[0]
This queues up a task on Deep Origin. When it completes, outputs will be written to the appropriate column in this database.
Multiple ligands¶
To run an end-to-end ABFE workflow on multiple ligands, we use:
job = sim.abfe.run_end_to_end(ligands=[sim.ligands[0],sim.ligand[1]])
Omitting the ligand IDs will run ABFE on all ligands in the Complex object.
jobs = sim.abfe.run_end_to_end()
Each ligand will be run in parallel on a separate instance.
Job Control¶
Visualize Job¶
Once a job has been submitted, you can track its status by inspecting the Job object:
job # returned from abfe.run_end_to_end()
Expected output
Watch jobs¶
To monitor the status of this job, use:
job.watch() # returned from abfe.run_end_to_end()
Stop watching jobs¶
To manually stop watching a job, do:
job.stop_watching() # returned from abfe.run_end_to_end()
Cancel jobs¶
To cancel a job, use:
job.cancel() # returned from abfe.run_end_to_end()
Parameters¶
Viewing parameters¶
The end to end ABFE tool has a number of user-accessible parameters. To view all parameters, use:
from deeporigin.drug_discovery import Complex, EXAMPLE_DATA_DIR
sim = Complex.from_dir(EXAMPLE_DATA_DIR)
sim.abfe._params.end_to_end
Expected output
This will print a dictionary of the parameters used for ABFE, similar to:
{
"abfe": {
"add_fep_repeats": 0,
"amend": "__NO_AMEND",
"annihilate": true,
"atom_mapping_threshold": 0.01,
"em_all": true,
"em_solvent": true,
"emeq_md_options": {
"T": 298.15,
"cutoff": 0.9,
"fourier_spacing": 0.12,
"hydrogen_mass": 2.0,
"Δt": 0.004
},
"fep_windows": [
{
"restraints_A": [
0.0,
0.01,
0.025,
0.05,
0.1,
0.35,
0.5,
0.75,
1.0
]
},
{
"coul_A": [
1.0,
0.8,
0.6,
0.4,
0.2,
0.0
]
},
{
"vdw_A": [
1.0,
0.9,
0.8,
0.7,
0.6,
0.5,
0.4,
0.3,
0.2,
0.1,
0.0
]
}
],
"mbar": 1,
"npt_reduce_restraints_ns": 2.0,
"nvt_heating_ns": 1.0,
"prod_md_options": {
"T": 298.15,
"barostat": "MonteCarloBarostat",
"barostat_exchange_interval": 500,
"cutoff": 0.9,
"fourier_spacing": 0.12,
"hydrogen_mass": 2.0,
"integrator": "BAOABIntegrator",
"Δt": 0.004
},
"repeats": 1,
"run_name": "binding",
"skip_emeq": "__NO",
"softcore_alpha": 0.5,
"steps": 1250000,
"system": "complex",
"test_run": 0,
"thread_pinning": 0,
"thread_pinning_offset": 0,
"threads": 0,
"workers": 0
},
"complex_prep": {
"include_ligands": 1,
"include_protein": 1,
"sysprep_params": {
"charge_method": "bcc",
"do_loop_modelling": false,
"force_field": "ff14SB",
"is_lig_protonated": true,
"is_protein_protonated": true,
"keep_waters": true,
"lig_force_field": "gaff2",
"ligand_res_names": [
"LIG"
],
"padding": 1.0,
"save_gmx_files": false
},
"test_run": 0,
"thread_pinning": 0,
"thread_pinning_offset": 0
},
"emeq": {
"amend": "__NO_AMEND",
"em_all": true,
"em_solvent": true,
"emeq_md_options": {
"T": 298.15,
"cutoff": 0.9,
"fourier_spacing": 0.12,
"hydrogen_mass": 2.0,
"Δt": 0.004
},
"from_run": "__USE_SYSTEM",
"npt_reduce_restraints_ns": 0.2,
"nvt_heating_ns": 0.1,
"test_run": 0,
"thread_pinning": 0,
"thread_pinning_offset": 0,
"threads": 0
},
"ligand_prep": {
"include_ligands": 1,
"include_protein": 0,
"sysprep_params": {
"charge_method": "bcc",
"do_loop_modelling": false,
"force_field": "ff14SB",
"is_lig_protonated": false,
"is_protein_protonated": false,
"keep_waters": false,
"lig_force_field": "gaff2",
"padding": 1.0,
"save_gmx_files": false
},
"test_run": 0,
"thread_pinning": 0,
"thread_pinning_offset": 0
},
"md": {
"amend": "__NO_AMEND",
"continue": 0,
"from_run": "__USE_SYSTEM",
"md_options": {
"T": 298.15,
"barostat": "MonteCarloBarostat",
"barostat_exchange_interval": 500,
"cutoff": 0.9,
"fourier_spacing": 0.12,
"hydrogen_mass": 2.0,
"integrator": "BAOABIntegrator",
"Δt": 0.004
},
"run_name": "md",
"steps": 250000,
"test_run": 0,
"thread_pinning": 0,
"thread_pinning_offset": 0,
"threads": 0
},
"solvation": {
"add_fep_repeats": 0,
"amend": "__NO_AMEND",
"annihilate": true,
"atom_mapping_threshold": 0.01,
"em_all": true,
"em_solvent": true,
"emeq_md_options": {
"T": 298.15,
"cutoff": 0.9,
"fourier_spacing": 0.12,
"hydrogen_mass": 2.0,
"Δt": 0.004
},
"fep_windows": [
{
"coul_A": [
1.0,
0.8,
0.6,
0.4,
0.2,
0.0
]
},
{
"vdw_A": [
1.0,
0.9,
0.8,
0.7,
0.6,
0.5,
0.4,
0.3,
0.2,
0.1,
0.0
]
}
],
"mbar": 1,
"npt_reduce_restraints_ns": 0.2,
"nvt_heating_ns": 0.1,
"prod_md_options": {
"T": 298.15,
"barostat": "MonteCarloBarostat",
"barostat_exchange_interval": 500,
"cutoff": 0.9,
"fourier_spacing": 0.12,
"hydrogen_mass": 2.0,
"integrator": "BAOABIntegrator",
"Δt": 0.004
},
"repeats": 1,
"skip_emeq": "__NO",
"softcore_alpha": 0.5,
"steps": 300000,
"test_run": 1,
"thread_pinning": 0,
"thread_pinning_offset": 0,
"threads": 0,
"workers": 0
}
}
Modifying parameters¶
Any of these parameters are modifiable using dot notation. For example, to change the number of steps in the MD step, we can use:
from deeporigin.drug_discovery import Complex, EXAMPLE_DATA_DIR
sim = Complex.from_dir(EXAMPLE_DATA_DIR)
sim.abfe._params.end_to_end.md.steps = 500000
Using test_run¶
The test run parameter can be used to run ABFE for a short number of steps, to verify that all steps execute without consuming too many CPU cycles. This should not be used to run production simulations.
To set the test run parameter to 1, we can use:
from deeporigin.drug_discovery import Complex, EXAMPLE_DATA_DIR
sim = Complex.from_dir(EXAMPLE_DATA_DIR)
sim.abfe.set_test_run(1)
Results¶
Viewing results¶
After initiating a run, we can view results using:
sim.abfe.show_results()
This shows a table similar to:
Expected output
Exporting results for analysis¶
These results can be exported for analysis using:
df = sim.abfe.get_results()
df
Expected output
| Binding | Solvation | AnalyticalCorr | Std | Total | ID | File | r_exp_dg | SMILES |
|---|---|---|---|---|---|---|---|---|
| 16.23 | -27.53 | -7.2 | 0.0 | -36.50 | Ligands-1 | brd-2.sdf | -9.59 | [H]C1=C([H])C(C(=O)N(C([H])([H])[H])C([H])([H])[H])=C([H])C(C2=C([H])N(C([H])([H])[H])C(=O)C3=C2C([H])=C([H])N3[H])=C1[H] |
| -454.99 | -722.01 | -7.58 | 0.0 | -259.44 | Ligands-2 | brd-3.sdf | -7.09 | [H]C([H])=C([H])C([H])([H])N1C(=O)C2=C(C([H])=C([H])N2[H])C(C2=C([H])C([H])=C([H])C(C(=O)N(C([H])([H])[H])C([H])([H])[H])=C2[H])=C1[H] |
| -600.31 | -1354.79 | -7.47 | 0.0 | -747.00 | Ligands-3 | brd-4.sdf | -8.64 | [H]C1=C([H])C(C(=O)N(C([H])([H])[H])C([H])([H])[H])=C([H])C(C2=C([H])N(C([H])([H])/C([H])=C([H])C([H])([H])[H])C(=O)C3=C2C([H])=C([H])N3[H])=C1[H] |