OpenDUck is an open-source implementation of dynamic undocking (DUck), an application of steered molecular dynamics developed as a fast algorithm to calculate the work necessary to reach a quasi-bound state, at which a certain native contact is broken between the ligand and receptor. Due to its cheap and fast nature, DUck is mainly employed as a post-docking filter during virtual screening campaigns.
More details can be found in the original dynamic undocking publication:
Ruiz-Carmona, S., Schmidtke, P., Luque, F. et al. Dynamic undocking and the quasi-bound state as tools for drug discovery. Nature Chem 9, 201–206 (2017). https://doi.org/10.1038/nchem.2660
We recommend you install OpenDUck into a fresh Conda environment.
$ git clone [email protected]:CBDD/openduck.git
$ cd openduck
$ conda env create -f environment.yaml
$ conda activate openduck
$ python setup.py install
OpenDUck can be used either as a Python library or as an executable. The executable provided is openduck, which is automatically added to your path if you install with pip or conda. Alternatively you can use openduck.py in the scripts subdirectory.
OpenDUck comes with several submodules to launch the different steps of dynamic undocking:
$ conda activate openduck
$ openduck -h
usage: openduck [-h] ...
Open-source toolkit for dynamic undocking.
optional arguments:
-h, --help show this help message and exit
Open-source dynamic undocking toolkit. Choose one of the following actions::
openmm-prepare Preparation of systems for OpenMM simulations.
openmm-full-protocol
OpenDUck full OpenMM protocol.
openmm-from-equilibrated
OpenDUck OpenMM protocol starting from a pre-equilibrated system.
openmm-from-amber OpenDUck OpenMM protocol starting from an AMBER topology and coordinates.
amber-prepare Preparation of systems, inputs and queue files for AMBER simulations.
report Generate a report for OpenDUck results.
chunk Chunk a protein for dynamic undocking.
Each of the OpenDUck modules accepts the input either as command line arguments or as a yaml file. The arguments in the input.yaml are expected to follow the command-line nomenclature.
The OpenMM implementation of OpenDUck is divided in two steps: the preparation using openmm_forcefields and the production (MD/SMD) following the dynamic undocking protocol.
During the preparation step, the protein is reduced to the pocket region during the chunking step, then the ligand and receptor are parametrized and the solvation box is generated. After equilibration, the sequential MD/SMD simulations, sampling and steering the main interaction from the specified distances, are run in OpenMM.
The openMM submodules are set up to allow independent usage of the previously mentioned steps:
* `openmm-full-protocol`: Executes the full OpenDuck protocol, from the protein-ligand files to the steered simulations
* `openmm-prepare`: Executes only the preparation and equilibration of the protein-ligand complex
* `openmm-from-equilibrated`: Executes only the production of MD/SMD runs from an equilibrated input.
* `openmm-from-amber`: Starts the openDuck simulations from an AMBER topology.
The complete OpenDUck protocol through OpenMM can be executed using the openmm-full-protocol submodule.
The full protocol executes the following steps:
* Chunking (optional)
* Preparation
* Ligand parametrization
* Receptor parametrization
* Box generation
* Equilibration
* (MD + SMD@300K + SMD@325K + WQB check ) x smd-cycles
$ openduck openmm-full-protocol -h
usage: openduck openmm-full-protocol [-h] [-y YAML_INPUT] [-l LIGAND] [-i INTERACTION] [-r RECEPTOR] [-g GPU_ID]
[--do-chunk] [-c CUTOFF] [-b] [-f {SMIRNOFF,GAFF2}] [-w {tip3p,spce}]
[-ff {amber99sb,amber14-all}] [-ion IONIC_STRENGTH]
[-s SOLVENT_BUFFER_DISTANCE] [-water WATERS_TO_RETAIN] [-fl]
[-F FORCE_CONSTANT_EQ] [-n SMD_CYCLES] [-m MD_LENGTH] [-W WQB_THRESHOLD]
[-v INIT_VELOCITIES] [-d INIT_DISTANCE]
Full dynamic undocking protocol in OpenMM. The ligand, receptor and solvation box are parameterized with the
specified parameters. If specified, the receptor is reduced to a chunked pocket. After equilibration, serial
iterations of MD and SMD are performed until the WQB or max_cycles threshold is reached.
optional arguments:
-h, --help show this help message and exit
Main arguments:
-y YAML_INPUT, --yaml-input YAML_INPUT
Input yaml file with all the arguments for the full OpenDUck protocol.
-l LIGAND, --ligand LIGAND
Ligand MOL file to use as reference for interaction.
-i INTERACTION, --interaction INTERACTION
Protein atom to use for ligand interaction.
-r RECEPTOR, --receptor RECEPTOR
Protein or chunked protein in PDB format used as receptor.
-g GPU_ID, --gpu-id GPU_ID
GPU ID; if not specified, runs on CPU only.
--keep-all-files Disable cleaning up intermediate files during simulations.
Chunking arguments:
--do-chunk Chunk initial receptor based on the interaction with ligand and add
cappings.
-c CUTOFF, --cutoff CUTOFF
Cutoff distance to define chunking (in Angstroms). Default = 9 A.
-b, --ignore-buffers Do not remove buffers (solvent, ions etc.)
Parameterization arguments:
-f {SMIRNOFF,GAFF2}, --small-molecule-forcefield {SMIRNOFF,GAFF2}
Small molecule forcefield to use for parameterization.
-w {tip3p,spce}, --water-model {tip3p,spce}
Water model to parameterize the solvent.
-ff {amber99sb,amber14-all}, --protein-forcefield {amber99sb,amber14-all}
Protein forcefield to parameterize the chunked protein.
-ion IONIC_STRENGTH, --ionic-strength IONIC_STRENGTH
Ionic strength (concentration) of the counter ion salts (Na+/Cl-). Default
= 0.1 M
-s SOLVENT_BUFFER_DISTANCE, --solvent-buffer-distance SOLVENT_BUFFER_DISTANCE
Buffer distance between the periodic box and the protein (in Angstroms).
Default = 10 A
-water WATERS_TO_RETAIN, --waters-to-retain WATERS_TO_RETAIN
PDB file containing structural water molecules to retain during
simulations. Default is waters_to_retain.pdb.
-fl, --fix-ligand Some simple fixes for the ligand: ensure tetravalent nitrogens have the
right charge assigned and add missing hydrogen atoms.
MD/SMD production arguments:
-F FORCE_CONSTANT_EQ, --force-constant-eq FORCE_CONSTANT_EQ
Force constant for equilibration.
-n SMD_CYCLES, --smd-cycles SMD_CYCLES
Number of MD/SMD cycles to perform.
-m MD_LENGTH, --md-length MD_LENGTH
Length of MD sampling between SMD runs in ns.
-W WQB_THRESHOLD, --wqb-threshold WQB_THRESHOLD
Minimum WQB threshold; if not reached after each SMD cycle, further
simulations will be terminated. If not set (default), all SMD cycles will
be run.
-v INIT_VELOCITIES, --init-velocities INIT_VELOCITIES
Set initial velocities when heating.
-d INIT_DISTANCE, --init-distance INIT_DISTANCE
Set initial hydrogen bond distance for SMD in Angstroms. Default = 2.5 A.
A valid example is provided in the test subfolder, and can be executed using the command-line arguments like the following:
$ openduck openmm-full-protocol -l brd4_lig.mol -r 4LR6_aligned_chunk_nowat.pdb -i A_ASN_140_ND2 \
-g 0 -f GAFF2 -ff amber14-all -w tip3p -w 6 -n 20
Similarly, an execution with the input parameters in a yaml is also possible (the example has slightly different parameters as the command line, just for ilustration purposes ).
openduck openmm-full-protocol -y input.yaml
where the input.yaml file has the following
# Main Arguments
interaction : A_ASN_140_ND2
receptor_pdb : 4LR6_aligned_chunk_nowat.pdb
ligand_mol : brd4_lig.mol
gpu_id : 0
# Chunking Arguments
do_chunk : True
cutoff : 10
ignore_buffers : False
# Preparation Arguments
small_molecule_forcefield : smirnoff
protein_forcefield : amber14-all
water_model : tip3p
ionic_strength : 0.05
waters_to_retain : waters_to_retain.pdb
solvent_buffer_distance : 15
force_constant_eq : 1
# Production Arguments
smd_cycles : 20
md_length : 0.5
wqb_threshold : 6
init_velocities : 0.00001
init_distance : 2.5
fix_ligand : False
The OpenDUck protocol can also be executed independently in two steps: the preparation & equilibration with the openmm-prepare subcommand and the production with the openmm-from-equilibrated subcommand. Following the schema presented above, each substep is performed by running the following subcommands.
openmm-prepare
* Chunking (optional)
* Preparation
* Ligand parametrization
* Receptor parametrization
* Box generation
* Equilibration
openmm-from-equilibrated
* (MD + SMD@300K + SMD@325K + WQB check ) x smd-cycles
Each of the two subcommads share arguments with the openmm-full-protocol, based on the execution steps they have in common. As with the previus command, the execution can come from flagged arguments or with the same arguments in a yaml input file.
Sample yaml input files for openmm-prepare and openmm-from-equilibrated are provided below:
# openmm-prepare input.yaml
# Main arguments
ligand_mol : '1a28_lig.mol'
receptor_pdb : '1a28_prot.pdb'
interaction : 'A_ARG_766_NH2'
# Chunking arguments
do_chunk : True
cutoff : 10
ignore_buffers : False
#Preparation arguments
small_molecule_forcefield : 'GAFF2'
protein_forcefield : 'amber14-all'
water_model : 'spce'
ionic_strength : 0.1
solvent_buffer_distance : 12
#Equilibration arguments
do_equilibrate : True
gpu_id : 0
force_constant_eq : 1
# openmm-from-equilibrated
equilibrated_system : equil.chk
pickle : complex_system.pickle
smd_cycles : 20
md_length : 0.5
wqb_threshold : 6
gpu_id : 0
Alternatively, the OpenDUck protocol can be launched from an AMBER topology and coordinates. This can be done using the openmm-from-amber subcommand.
$ openduck openmm-from-amber -h
usage: openduck openmm-from-amber [-h] [-y YAML_INPUT] [-c COORDINATES] [-t TOPOLOGY] [-i INTERACTION]
[-r RECEPTOR] [-n SMD_CYCLES] [-m MD_LENGTH] [-W WQB_THRESHOLD]
[-v INIT_VELOCITIES] [-d INIT_DISTANCE] [-g GPU_ID]
OpenDUck OpenMM protocol starting from an AMBER topology and coordinates. Using the AMBER topology (prmtop) and
coordinates (inpcrd), identifies the main interaction and serial iterations of MD and SMD are performed until the
WQB or max_cycles threshold is reached.
optional arguments:
-h, --help show this help message and exit
-y YAML_INPUT, --yaml-input YAML_INPUT
Input yaml file with all the arguments for the OpenMM simulations from the
equilibrated system.
-c COORDINATES, --coordinates COORDINATES
AMBER input coordinates.
-t TOPOLOGY, --topology TOPOLOGY
AMBER input topology.
-i INTERACTION, --interaction INTERACTION
Protein atom to use for ligand interaction.
-r RECEPTOR, --receptor RECEPTOR
Receptor MOL2 file.
-n SMD_CYCLES, --smd-cycles SMD_CYCLES
Number of MD/SMD cycles to perform.
-m MD_LENGTH, --md-length MD_LENGTH
Length of MD sampling between SMD runs in ns.
-W WQB_THRESHOLD, --wqb-threshold WQB_THRESHOLD
Minimum WQB threshold; if not reached after each SMD cycle, further
simulations will be terminated. If not set (default), all SMD cycles will
be run.
-v INIT_VELOCITIES, --init-velocities INIT_VELOCITIES
Set initial velocities when heating.
-d INIT_DISTANCE, --init-distance INIT_DISTANCE
Set initial hydrogen bond distance for SMD in Angstroms. Default = 2.5 A.
-g GPU_ID, --gpu-id GPU_ID
GPU ID; if not specified, runs on CPU only.
--keep-all-files Disable cleaning up intermediate files during simulations.
If prefered, the script can also be executed from a yaml file with the commands. A sample input.yaml file could be the following:
interaction: A_ASN_140_ND2
coordinates: system_complex.inpcrd
topology: system_complex.prmtop
receptor: 4LR6_aligned_chunk_nowat.pdb
smd_cycles: 20
wqb_threshold: 6
gpu_id: 0
Originally dynamic undocking was designed to run in AMBER; in OpenDUck we have maintained the same protocol for AMBER, replacing the parametrization and file-generation previously done by MOE with an open-source implementation. As with the original protocol, the user is free to choose their preferred queueing headers and preparation via either the commandline arguments or the input yaml file. Preparation flags are very similar to those for OpenMM, as most of the implementation is shared.
$ openduck amber-prepare -h
usage: openduck amber-prepare [-h] [-y YAML_INPUT] [-l LIGAND] [-i INTERACTION] [-r RECEPTOR] [--do-chunk]
[-c CUTOFF] [-b] [-f {SMIRNOFF,GAFF2}] [-w {tip3p,spce,tip4pew}]
[-q QUEUE_TEMPLATE] [-H] [-n SMD_CYCLES] [-W WQB_THRESHOLD]
[-ff {amber99sb,amber14-all}] [-ion IONIC_STRENGTH] [-s SOLVENT_BUFFER_DISTANCE]
[-water WATERS_TO_RETAIN] [--seed SEED] [-B] [-t THREADS] [-fl]
Preparation of systems, inputs and queue files for AMBER simulations. The ligand, receptor and solvation box are
parameterized with the specified parameters. If specified, the receptor is reduced to a chunked pocket for a
faster production. The input and queue files are prepared from templates found in the duck/templates directory.
optional arguments:
-h, --help show this help message and exit
Main arguments:
-y YAML_INPUT, --yaml-input YAML_INPUT
Input yaml file with all the arguments for the system preparation and
inputs/queueing for AMBER.
-l LIGAND, --ligand LIGAND
Ligand MOL file to use as reference for interaction.
-i INTERACTION, --interaction INTERACTION
Protein atom to use for ligand interaction.
-r RECEPTOR, --receptor RECEPTOR
Protein or chunked protein in PDB format used as receptor.
Chunking arguments:
--do-chunk Chunk initial receptor based on the interaction with ligand and add
cappings.
-c CUTOFF, --cutoff CUTOFF
Cutoff distance to define chunking (in Angstroms). Default = 9 A.
-b, --ignore-buffers Do not remove buffers (solvent, ions etc.)
Parameterization arguments:
-f {SMIRNOFF,GAFF2}, --small-molecule-forcefield {SMIRNOFF,GAFF2}
Small molecule forcefield to use for parameterization.
-w {tip3p,spce,tip4pew}, --water-model {tip3p,spce,tip4pew}
Water model to parameterize the solvent.
-q QUEUE_TEMPLATE, --queue-template QUEUE_TEMPLATE
Write out a queue file from templates.
-H, --HMR Perform Hydrogen Mass Repartition on the topology and use it for the input
files.
-n SMD_CYCLES, --smd-cycles SMD_CYCLES
Number of SMD replicas to perform.
-W WQB_THRESHOLD, --wqb-threshold WQB_THRESHOLD
Minimum WQB threshold; if not reached after each SMD cycle, further
simulations will be terminated. If not set (default), all SMD cycles will
be run.
-ff {amber99sb,amber14-all}, --protein-forcefield {amber99sb,amber14-all}
Protein forcefield to parameterize the chunked protein.
-ion IONIC_STRENGTH, --ionic-strength IONIC_STRENGTH
Ionic strength (concentration) of the counter ion salts (Na+/Cl-). Default
= 0.1 M
-s SOLVENT_BUFFER_DISTANCE, --solvent-buffer-distance SOLVENT_BUFFER_DISTANCE
Buffer distance between the periodic box and the protein (in Angstroms).
Default = 10 A
-water WATERS_TO_RETAIN, --waters-to-retain WATERS_TO_RETAIN
PDB file containing structural water molecules to retain during
simulations. Default is waters_to_retain.pdb.
--seed SEED Specify seed for AMBER inputs.
-fl, --fix-ligand Some simple fixes for the ligand: ensure tetravalent nitrogens have the
right charge assigned and add missing hydrogen atoms.
Batch argments:
-B, --batch Enable batch processing for multi-ligand SDF.
-t THREADS, --threads THREADS
Define number of CPUs for batch processing.
--keep-all-files Disable cleaning up intermediate files during preparation and simulations.
--resume Enable the resume mode. Protecting LIG_target folders already prepared and
starting from the not done, avoiding overwritting.
-i0 INDEX0, --index0 INDEX0
Starting index for naming batch ligands. Default: 1.
-p PREFIX, --prefix PREFIX
Prefix to name ligand folder during batch preparation. Default:
LIG_target.
$ openduck amber-prepare -l ../1a28_lig.mol -r ../1a28_prot.pdb -i A_ARG_766_NH2 \
--do-chunk -c 10 -b False \
-f GAFF2 -ff amber99sb -w SPCE -ion 0.05 -s 11 \
--HMR -n 4 -W 4 -q Slurm
Typically, dynamic undocking is employed as a post-docking filter in a virtual screening campaign. As such, multiple ligands might be generated simultaneously. A batch execution of the DUck preparation can be executed specifying the --batch argument and providing an SD-file with multiple ligands as input. Each ligand and the associated file structure will be generated in the LIG_target_{n} subfolder, where n is the ligand's position in the input SDF.
$ openduck amber-prepare -l ../brd4_ligands.sdf -r ../4LR6_aligned_chunk_nowat.pdb -i A_ASN_140_ND2 \
--waters-to-retain ../waters_to_retain.pdb -f gaff2 -ff amber14-all -w spce -ion 1 -s 30
--HMR --smd-cycles 10 -wqb-threshold 8 -q SGE --seed -1 --batch --threads 8
The queueing template is stored in duck/templates/queueing_templates. To customize a template, a new file can be added in the directory with the following format: {queue_name}_array.q or {queue_name}.q either if the expected execution is in an array or not. For a local execution, the local argument can be given for a plain list of commands.
The templates have the following format:
#!/bin/bash
#HEADER
{functions}
{commands}
After successfully running OpenDUck using either OpenMM or AMBER, the results can be compiled using the report submodule.
By default, the lowest
Important
Due to the different formating of AMBER and OpenMM steered molecular dynamics output, the --format argument is needed when analyzing either output.
When analyzing a single result, the default pattern is the current directory (.). If you wish to analyze multiple directories, specify the wildcard pattern.
$ openduck report -h
usage: openduck report [-h] [-p PATTERN] [-d {min,single,avg,jarzynski,all}] [-o OUTPUT] [-of {csv,sdf,tbl}]
[--plot] [-s SUBSAMPLE_SIZE] [-i ITERATIONS] [-t STEP_THRESHOLD] [-f {amber,openmm}]
Generate a table report for dynamic undocking output. For a multi-ligand report, use the pattern flag with
wildcards to the directories.
optional arguments:
-h, --help show this help message and exit
-p PATTERN, --pattern PATTERN
Wildcard pattern to find folders with DUck data.
-d {min,single,avg,jarzynski,all}, --data {min,single,avg,jarzynski,all}
Mode to compile the report [min | single | avg | jarzynski | all]
-o OUTPUT, --output OUTPUT
Output file; default is printing report to stdout.
-of {csv,sdf,tbl}, --output-format {csv,sdf,tbl}
Output format, [csv | sdf | tbl].
--plot Plot work or energy values to file.
-s SUBSAMPLE_SIZE, --subsample-size SUBSAMPLE_SIZE
Subsample size for Jarzynski bootstrapping.
-i ITERATIONS, --iterations ITERATIONS
Number of bootstrapping iterations for Jarzynski analysis.
-t STEP_THRESHOLD, --step-threshold STEP_THRESHOLD
Step threshold to find the minima. Only needed with custom executions of DUck. Default =
2500 steps
-f {amber,openmm}, --format {amber,openmm}
Engine used to generate results. Default = amber
Parsing
$ cd Iridium/1a28/ARG766_HN2
$ openduck report -p '*/'
System WQB
SMIRNOFF/ 5.6297999999999995
GAFF2/ 5.74429
SMIRNOFF_nohmr/ 6.922040000000001
GAFF2_nohmr/ 6.264200000000001
GAFF2_bigbox/ 5.180540000000001
GAFF2_tip4pew/ 7.60925
GAFF2_spce/ 6.17347
Calculating and reporting the
$ openduck report -p '*/' -d jarzynski
System Jarzynski Jarzynski_SD Jarzynski_SEM
SMIRNOFF/ 6.391165243930965 0.12701922010484387 0.02033935321315065
GAFF2/ 6.990007044361647 0.23871054814271678 0.038224279367893604
SMIRNOFF_nohmr/ 7.690209327251983 0.10126872662893496 0.01621597423328342
GAFF2_nohmr/ 7.075924262571702 0.10628038471267455 0.01701848178973498
GAFF2_bigbox/ 6.670905906961162 0.3167545344338863 0.05072131880828811
GAFF2_tip4pew/ 9.007291835327141 0.33968295997031067 0.05439280525909316
GAFF2_spce/ 7.31123937991228 0.24521160219077848 0.03926528115039682