README

This repository, stores the input files necessary to execute the gromacs/coco workflow for a penta-alanine example in a High Performance Computing (HPC) environment such as STAMPEDE (XSEDE, US supercomputing resources) or ARCHER (UK supercomputing resources).

The resource configuration file stampede.rcfg and archer.rcfg should be modified by the user, with their credentials on the username UNAME and allocation (project to be charged for the computations on the HPC resource).

In addition the users would ideally use their own simulation files, for computation contexts similar to this one and complete the information on their file names in a workload configuration file such as gmxcoco.wcfg.

A- In this case, we will perform a set of Molecular Dynamics (MD) simulations (the user decides the size of the set throught the num_CUs parameter of the gmxcoco.wcfg configuration file) with gromacs, starting from an initial structure ./inp_files/helix.gro with simulation parameters and topology described at respectively ./inp_files/grompp-verlet.mdp and ./inp_files/topol.top

B- Afterwards, we will analyse all coordinate trajectories produced from the set of MD simulations by means of the CoCo method through the pyCoCo tool, and produce new simulation starting points (in terms of .gro structures) by means of the same technique.

C- We will then perform a set of minimisation procedures using position restraints and the simulation parameters described at ./inp_files/grompp_after_coco-em1.mdp, ./inp_files/grompp_after_coco-em2.mdp Afterwards, we will proceed with a standard set of MD simulation runs using the same simulation parameters as described at step A.

D- We will analyse all coordinate trajectories produced from the set of MD simulations of step C by means of the CoCo method as described for step B.

Steps from A to D make 2 iterations in our workflow:

• Iteration 1 - Step A & Step B
• Iteration 2 - Step C & Step D

All iterations from the 2nd onwards would include one C-like step and one D-like step for example:

• Iteration 3 - Step C & Step D
• Iteration 4 - Step C & Step D

How do I get the workflow to run?

To run this workflow in the workflow machine the users should do the following:

#!python

ssh [email protected]

wget https://github.com/ashkurti/Hands-on-session-part-2-The-customized-gromacs-coco-workflow-of-ExTASY/archive/ext_gmxcoco-0.7.tar.gz  
tar -xvf ext_gmxcoco-0.7.tar.gz
cd Hands-on-session-part-2-The-customized-gromacs-coco-workflow-of-ExTASY-ext_gmxcoco-0.7

source /opt/tutorials/radical-tutorial/rp-tut.sh
virtualenv --system-site-packages $HOME/ve
source $HOME/ve/bin/activate
pip install radical.ensemblemd
export RADICAL_ENMD_VERBOSE=REPORT

and at last, after completing stampede.rcfg and gmxcoco.wcfg appropriately:

#!python

python extasy_gmxcoco.py --RPconfig stampede.rcfg --Kconfig gmxcoco.wcfg

If everything is successful, with the current configuration the user should notice the generation of an output folder in his/her current directory with the following contents:

#!python

ls -l output/*
-rw------- 1 uname pa  491 Oct 20 12:05 output/coco-iter0.log
-rw------- 1 uname pa  538 Oct 20 12:16 output/coco-iter1.log

output/iter0:
total 3424
-rw------- 1 uname pa 1752276 Oct 20 12:05 md-0_0_whole.xtc
-rw------- 1 uname pa 1751064 Oct 20 12:05 md-0_1_whole.xtc

output/iter1:
total 3424
-rw------- 1 uname pa 1751604 Oct 20 12:15 md-1_0_whole.xtc
-rw------- 1 uname pa 1751592 Oct 20 12:15 md-1_1_whole.xtc

Who do I talk to?

• [email protected] (for the development of the standalone simulation workflow, development of the kernels to implement this workflow in an extasy context, input/output results expected at each step)
• [email protected] (for the input/output results expected at each step of the workflow and the simulation trajectories analysis technique used)
• [email protected] (for doubts on the interaction of the kernels with the underlying cyber tools - such as radical pilot, saga-python and radical-utils of the RADICAL team at Rutgers)
• [email protected] (for the cyber tools - such as radical pilot, saga-python and radical-utils of the RADICAL team at Rutgers)
• [email protected], [email protected], [email protected] (for the gromacs input files - *mdp, *gro, *top)