Switch branches/tags
Nothing to show
Find file History
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.




    Molecular dynamics (MD) simulations are electronic microscopes that allow researchers to
    study atomic details of biological systems (1). Applying Newton’s laws of motion, MD
    simulations can predict atom movements following a time series. MD simulations usually
    involve four steps: minimization, heating, equilibration, and production, and the purpose 
    of this Amber GPU MD simulation workflow is to automate these processes using graphics
    processing units (GPUs). This workflow can be used for all-atom molecular dynamic
    simulations, which involve five steps of minimization, one step of heating, three steps of
    equilibration, and one or more instances of production. MD simulations simulate atomic
    movements to help scientists predict reaction mechanisms, identify druggable pockets, and
    understand allosteric regulation (1). Moreover, due to the simple nature of this workflow,
    the workflow can be used as a starting point for learning how to run MD simulations. It 
    can also be used a starting point developing workflows that involve more complicated 
    aspects MD. It is our hope that this workflow will enable improved dissemination and 
    reproducibility in the field of molecular dynamics.

    ExecutionChoice    This parameter gives the option to run same workflow either on a local 
		       machine or in a GPU cluster.
    NP   	       The number of CPU processors for the minimization steps. It is the same 
 		       number of processors for each of the step.
    CMPD 	       The folder name containing the topology and coordinate files. The folder 
		       name needs to match the stem name of the topology and coordinate file.
    AMBERHOME_local    $AMBERHOME/bin on your local machine. This should be the path that points 
		       to where pmemd.MPI is located.
    inputFolder        The inputFolder path on your local machine. It is the directory where the
		       confDir and CMPD folder is located. An absolute path is required.
    TargetHost         The remote cluster login.
    IdentityFile       The identity file (id_rsa) that allows remote access. A complete and 
		       absolute path is required.
    remoteDir          Work Directory on remote GPU cluster. A complete and absolute path is 
    numOfJobs  	       This parameter will allow "n" production jobs to be submitted in parallel 
		       for a single system. This parameter is only for job submissions in GPU
    AMBERHOME_cpu      $AMBERHOME on remote cluster on CPU nodes.
    AMBERHOME_gpu      $AMBERHOME on remote cluster on GPU nodes.
    cpuModulesToLoad   Modules dynamically modifies a user’s execution environment to configure 
		       the shell for an application.
    gpuModulesToLoad   The parameter “gpuModulesToload” prepares execution environment on GPU 
                       nodes on a cluster.
    t-cpu	       Time requests for each of the minimization step in the remote cluster.
    t-gpu	       Time requests for the heating step, equilibration steps and production step 
		       in the remote cluster.

Input & Output
    The input for this workflow is a set of directories include the MD simulation input scripts,
    system topology and coordinate files. Output files are list of plots, simulation trajectories,
    intermediate files, restart files, etc. Sample inputs and outputs are available and included 
    as a part of the download. They are in the inputs and outputs folders, respectively.

Operating System Requirment

    An extensive user guide/manual comes along with this workflow. For any question, please 
    consult the manual. Also, there is an appendix section at the end of the maunaul where it
    answer some of the questions users may have if the workflow fails to run.