Elastic Network Model

Double-well Elastic Network model of NtrC based on the parameterization in

E. Vanden-Eijnden and M. Venturoli, “Revisiting the finite temperature string method for the calculation of reaction tubes and free energies,” J Chem Phys 130, 194103 (2009).

Running Weighted Ensemble simulations

To run the Weighted Ensemble simulation, use the script, bin/run_we.py

$ python bin/run_we.py -h
usage: run_we.py [-h] -c CONFIG_FILE -n NAME [-p [PROTOCOLS [PROTOCOLS ...]]]
                 [-w NWORKERS] [--profile] [--timeout TIMEOUT]

WEST run script

optional arguments:
  -h, --help            show this help message and exit
  -c CONFIG_FILE        yaml config file name
  -n NAME               simulation name to run
  -p [PROTOCOLS [PROTOCOLS ...]]
                        protocols to run; by default run all
  -w NWORKERS           number of cores to use
  --timeout TIMEOUT     timeout when running with zmq work manager
  --profile             Profile code
  --no-run              Only setup simulations but do not run them

The -p flag specifies which protocols to run (DEFAULT: run all), and --no-run sets up the simulation directorie and scripts, but does not launch the actual simulation. The protocols are specified in the yaml configuration file in configs and the are selected by the name variable. When running with more than one worker, the simulation is run with the zmq work manager, and the --timeout flag should be set. Since the re-weighting step can be expensive the timeout parameter should be set between 1800 and 3600 seconds, although this will depend on the number of cores available for the calculation and the processor speed.

The command used for the paper was:

$ python bin/run_we.py -c configs/we_run_config_freestr.yaml -n we_freestr -w 6 --timeout 3600

Since this calculation can take some time, nohup should probably be used.

Note

The propagator_block_size in we_run_config_freestr.yaml, which sets the block_size parameter in the configuration file, should be adjusted so that (nbins * target_count) / propagator_block_size >= nworkers to ensure proper scaling of the simulation when varying the number of workers or the total number of replicas when the bin space is fully populated.

Analyzing data

To analyze the Weighted Ensemble simulations and calculate the probability distribution based on the final string configuration, use the script, analysis/get_bin_populations_all.py

$WEST_ROOT/bin/west analysis/get_bin_populations_all.py -h
Entering Python shell using python2.7
-----------------------
get_bin_populations_all.py
-----------------------
usage: get_bin_populations_all [-h] [-r RCFILE]
                               [--quiet | --verbose | --debug] [--version]
                               [-o H5OUT]

Retrieve strings from west.h5 file and write them to new file

optional arguments:
  -h, --help            show this help message and exit
  -o H5OUT              name of output file

general options:
  -r RCFILE, --rcfile RCFILE
                        use RCFILE as the WEST run-time configuration file
                        (default: west.cfg)

Run this script as:

$ cd we_freestr
$ $WEST_ROOT/bin/west ../analysis/get_bin_populations_all.py -o bin_populations_all.h5 -r we_phase1.cfg

Generating figures

To generate the figures that appear in the manuscript after running all of the simulations and the scripts that analyze them, run:

$ cd generate_figures
$ python plot_pmf.py

This will create one .eps file per figure.