/mmgbsa

Primary LanguageRich Text Format

Table of contents:

Setup Environment

This implementation of molecular mechanics-Generalized Born Surface Area (MM-GBSA) has several software dependencies:

  • VMD (v1.9.3)
  • CHARMM (Free Version 43b2)
  • Perl (v5.16.3)
  • Slurm (v17.02.7) # job scheduler

This code was built and tested against the software versions described above, but using other similar versions is probably fine (caveat emptor). Please consult each of the above's website for installation documentation.

Calculation Setup

The first step in using these tools is to first clone a copy of this repository, in a directory that is appropriate for running this calculation:

cd wherever_you_wish_to_run
git clone git@github.com:weinsteinlab/mmgbsa.git

Optionally, you can rename the cloned repository to something meaningful for your calculations

mv mmgbsa my_mmgbsa_run

Once this is done, go into this directory:

cd my_mmgbsa_run # or whatever the directory is named at this point

Note: directory is hereafter assumed to be your current working directory ./

Editing ./scripts/setenv.sh

First, you will need to edit ./scripts/setenv.sh, in which you must provide paths to the software dependencies described above. Specifically, you'll need to:

  1. Edit: export mmgbsa_path=/absolute/path/to/parent/mmgbsa # this must be the absolute path to your MM-GBSA parent directory.

    Next, you'll see 2 if blocks. This gives you the ability to set different paths to software dependencies for different clusters/file systems.

  2. For each if block, you'll see several [[ $HOSTNAME = cluster_node_name ]] Simply change cluster_node_name to whatever your cluster uses for compute node names. The first if block has 3 of these names, but you really only need to edit one of them (you can keep the extraneous [[]] or delete them-it doesn't make a difference).

  3. Within each if block, you'll need to change the absolute path to CHARMM, VMD, and Perl.

Editing ./RUN.sh

  1. Edit mmgbsa_path. This must be the absolute directory path to the parent MM-GBSA directory.

  2. You will also need to define the following variables that define the system, interacting partners, input files, etc. (each is described in ./RUN.sh):
    • system_selection # should only include two protein/molecules of interest
    • partA_selection= # first interacting protein/molecule
    • partB_selection= # first interacting protein/molecule
    • cutoff_residues= # controls which residues are included in decomposition, unless manual definitions are provided (via the next to variables described)
    • partA_residues_selection= # Part A residues to include in per-residue energy decomposition
    • partB_residues_selection= # Part B residues to include in per-residue energy decomposition
    • traj= # absolute path to your trajectory file--.dcd or . xtc can be used.

  3. You will also need to edit the following variables that control job parallelization:
    • start_frame= # first frame in the dcd file to use (index-1)
    • frame_stride= # e.g., if set to 2, only use every other trajectory frame
    • frames_per_job= # number of frames per subjob. The smaller this number is, the more subjobs that will be run (where each subjob runs on its own CPU core)
    • n_jobs= # how many subjobs. Should equal ceiling(frames/frames_per_job)
    • max_jobs_running_simultaneously= # consider limiting the number of concurrently running subjobs if on a shared cluster.
    • queueing_system= # should be "SLURM" (most robustly tested), though "LSF" and "SGE" are implemented.
    • queue_name= # Slurm partition to use for subjobs

  4. More variables related to the model investigated:
    • proteins= # number of proteins in the system
    • For each protein, you'll need the following lines: capping[1]= # is the first protein/peptide's n-term capped? 0 for no, 1 for yes. capping[2]= # is the first protein/peptide's n-term capped? 0 for no, 1 for yes.

      the next protein would continue incrementing the capping number: capping[3], capping[4], ...

Editing ./top_files_mackerell

  • The directory ./top_files_mackerell should include all of the ./.prm, ./.rtf, and ./.str files used to run the simulation.

Adding input files

  1. Create the directory input with the following command in the parent directory mkdir input
  2. Add the following files to `input:
    • your system's .psf file
    • a .pdb of the first frame of the trajectory

Running MM-GBSA

After performing all of the above edits, you're ready to submit the job! I would recommend launching the job from an interactive job session (i.e., not the login/submission node), as the job submission has brief interludes of non-trivial computation.

Submitting the job

To submit the job, first run the following command, which does some pre-job setup (in the parent directory): ./scripts/generate_loader_str.sh # note, this command is NOT submitted with sbatch!

Once the prior command returns, run the following command in the parent directory: ./RUN.sh # note, this command is NOT submitted with sbatch! Follow on-screen prompts and instructions. This script will setup and submit all of the needed subjobs.

Running this script will take anywhere from 2-10 minutes.

Check job status

The status of the MD swarm can be checked with the following command:

squeue -u your_username