User Provided Endstate Simulation

Running host-guest (CB7) absolute free energy calculation example:

If the user wishes to run there own endstate simulation they just need to provide there own .parm7 and .nc files. Please remove all solvent molecules from the .parm7 and .nc files if used prior. As-well make sure the .parm7 file has the recommened bondi radii for specified GB solvent model (check AMBER manual). All the example inputs scripts can be found in script_examples directrory. Lets first take a look at our example YAML config file config_script1.yaml.

hardware_parameters:
    working_directory: working_directory
    executable: "pmemd.MPI" # executable machine for MD choice : [sander, sander.MPI, pmemd, pmemd.MPI, pmeded.CUDA]
    mpi_command: srun # system dependent /
    CUDA: False # run on GPU
    output_directory_name: basic_md_dir
    cache_directory_output: path/to/cache_directory
    top_directory_name: basic_md_dir

endstate_parameter_files:
    complex_parameter_filename: script_examples/structs/cb7-mol01_Hmass.parm7 # path to topology file; ["path/to/complex.parm7"]
    complex_coordinate_filename: script_examples/structs/cb7-mol01_user_provided_traj.nc # path to coordinate ["path/to/complex.ncrst"]list of coordinate file of a complex
    ligand_parameter_filename:  script_examples/structs/M01.parm7
    ligand_coordinate_filename: script_examples/structs/M01_user_provided_traj.nc
    receptor_parameter_filename: script_examples/structs/CB7.parm7
    receptor_coordinate_filename: script_examples/structs/CB7_user_provided_traj.nc

number_of_cores_per_system:
    complex_ncores: 2 #total number of cores per job
    ligand_ncores: 1 #total number of cores per ligand simulation
    receptor_ncores: 1 #total number of cores per ligand simulation

AMBER_masks:
    receptor_mask: ":CB7" # list of Amber masks denoting receptor atoms in respected complex file
    ligand_mask: ":M01" # list of Amber masks denoting ligand atoms in respected complex file

workflow:
    endstate_method: 0 #options REMD or 0 (meaning no endstate simulation will be performed just intermidates)endstate_method: REMD #options REMD, MD or 0 (meaning no endstate simulation will be performed just intermidates)
    intermediate_states_arguments:
        mdin_intermidate_config: script_examples/user_intermidate_mdin_args.yaml #intermidate mdins required states 3-8
        igb_solvent: 2 #igb [1,2,3,7,8]
        temperature: 300
        exponent_conformational_forces: [
            -8, -7, -6, -5, -4, -3, -2, -1, 0, 1, 2,
            2.584963, 3, 3.584963, 4
        ]  # list exponent values 2**p
        exponent_orientational_forces: [
            -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6,
            6.584963, 7, 7.584963, 8
        ]  # list exponent values 2**p
        restraint_type: 2 # choices: [ 1: CoM-CoM, 2: CoM-Heavy_Atom, 3: Heavy_Atom-Heavy_Atom, must be 1, 2 or 3 ]

The endstate_method is set to 0 which assumes the user provided there own endstate simulation. Please note if endstate_method is set to 0 the user must provided the following keys values for ligand_parameter_filename, ligand_coordinate_filename, receptor_parameter_filename, and receptor_coordinate_filename. Every key within the YAML config file has been covered in previous section YAML Config Syntax.

Running entire workflow

The recommend way for running this workflow is by submitting an batch file. This ensures the correct number of resrources and the submission of a single job for entire workflow. Example of batch file for SLURM schedular:

#!/bin/bash
#SBATCH --partition=main
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=12  # Use high number for post-energy analysis (imin=5) parallelization
#SBATCH --time=09:30:00
#SBATCH --job-name=ex_01
#SBATCH --export=all

pwd

mkdir /scratch/username/

run_implicit_ddm.py file:jobstore_example_03 --config_file script_examples/config_files/user_defined_endstate.yaml --workDir /scratch/username/

Performance Considerations

High CPU Count: Use a high number of CPUs (12+ recommended) for optimal performance. The Sander post-energy analysis (imin=5) runs on CPUs and completion time scales linearly with the number of CPUs - more CPUs = faster completion.