Running OpenMMDL Simulation

This page details the variables required to run the simulation and showcases the application of OpenMMDL Simulation.

OpenMMDL Simulation starts the MD simulation with the inputs acquired from OpenMMDL Setup.

This page details the variables that are required for starting the script and showcases the application of OpenMMDL Simulation.

Variables

OpenMMDL Simulation consists of mandatory and optional variables. The following are listed down below:

Mandatory:

-f = name of the folder, where the MD simulation files are stored
-t = Topology file of your protein from the setup including the path
-s = Python script from the setup including the path

Optional:

-l = Ligand/cofactor/additional-molecule file(s) in SDF, MOL or MOL2 format. Repeat -l or pass multiple files after one -l.
-c = Coordinates file of Amber
--failure-retries = Number of reruns if OpenMM fails with "Particle coordinate is NaN"

Application

An example of how a command line of OpenMMDL Simulation should look is:

openmmdl simulation -f {path/to/folder_name} -t {path/to/topology} -s {path/to/script} -l {path/to/ligand}

For systems with multiple ligands or cofactors, -l can be repeated or followed by multiple files in one call.

openmmdl simulation -f {path/to/folder_name} -t {path/to/topology} -s {path/to/script} -l {path/to/ligand1} {path/to/ligand2}

If a simulation exits with the OpenMM error Particle coordinate is NaN, OpenMMDL Simulation can automatically retry the run while keeping the original input files, with the default being 5 retries.

For help during the usage of OpenMMDL Simulation use the following command line:

openmmdl simulation -h all

Running OpenMMDL Simulation test simulations

There are two Systems prepared for the testing of the simulation.

1: A 10ns simulation of the 6B73 Protein-ligand complex with POPC membrane and TIP3P-FB water. To run the testing of 6b73 enter the following command line:

openmmdl simulation -f 6b73_testing_simulation -t ~/OpenMMDL/openmmdl_simulation/testing_sytems/6b73_membrane/6b73-moe-processed_openMMDL.pdb -s ~/OpenMMDL/openmmdl_simulation/testing_sytems/6b73_membrane/6b73_simulation.py -l  ~/OpenMMDL/openmmdl_simulation/testing_sytems/6b73_membrane/6b73_lig.sdf

2: A 10ns simulation of the 5WYZ Protein-ligand complex with TIP3P water. To run the testing of 5WYZ enter the following command line:

openmmdl simulation -f 5wyz_testing_simulation -t ~/OpenMMDL/openmmdl_simulation/testing_sytems/5wyz_solvent/5wyz-moe-processed_openMMDL.pdb -s ~/OpenMMDL/openmmdl_simulation/testing_sytems/5wyz_solvent/5wyz_simulation.py -l  ~/OpenMMDL/openmmdl_simulation/testing_sytems/5wyz_solvent/5VF.sdf

Each of the command lines should generate a folder where the script and input data are copied before running the MD simulation and postprocessing of the system.

The generated simulation scripts now include an equilibration stage before the production simulation. The final postprocessing layout depends on the selected MDAnalysis output mode and may create either Final_Output/All_Atoms, Final_Output/Prot_Lig, or both.

Running OpenMMDL Simulation using slurm

Two scripts are needed to run simulations via slurm. Start using the runOpenMM_slurm.sh bash script when being in the repository folder. It has several inputs. For help just type:

bash runOpenMM_slurm.sh

without any flags, it will list the flags needed. The simplest way to run a simulation is to use the -i flag, which takes an input directory including the simulation script, the topology and optionally the ligand file and it will create the outputs folder within the given directory. NOTE: make sure only one topology is present in the input folder so that it finds it automatically.

The script calls a second script (you don’t need to do that) that is used for slurms sbatch command to run multiple replicas. The second script can be left where it is and named how it is (SlurmWrap.sh).

One example line of how to start five replicas on cn-gpus would be:

bash runOpenMM.sh -i /path/to/Input/Folder -n 5 -c