Reactive flux in OpenPathSampling

Software Technical Information

The information in this section describes OpenPathSampling as a whole. Information specific to the additions in this module are in subsequent sections.

Language
Python (2.7)
Documentation Tool
Sphinx, numpydoc format (ReST)
Application Documentation
http://openpathsampling.org
Relevant Training Material
http://openpathsampling.org/latest/examples/
Licence
LGPL, v. 2.1 or later

Authors: Andreas Singraber

This module implements the reactive flux method in OpenPathSampling.

Purpose of Module

The reactive flux method in combination with a free energy calculation allows to derive the rate constant of a rare event. This is accomplished by a shooting algorithm similar to a committor analysis where fleeting trajectories starting from the dividing surface are generated and statistics about their state with respect to a collective variable is collected. There are many flavors of the reactive flux method, this module implements the effective positive flux method as described by van Erp and Bolhuis (see e.g. http://dx.doi.org/10.1016/j.jcp.2004.11.003).

The implementation introduces the following new classes:

  • ReactiveFluxSimulation inherits from ShootFromSnapshotsSimulation and implements the shooting algorithm similar to CommittorSimulation. First, backward trajectories from the initial snapshots are started and followed until they either hit state A or recross the dividing surface. In the latter case the trajectory is rejected. If instead the trajectory reaches A, a forward shot is performed until the trajectory reaches either A (rejected) or B (accepted). The forward trajectory is allowed to recross the barrier any number of times but must end up in B without reaching A. To implement this behaviour of a forward shot depending on the final state of the backward trajectory a NonCanonicalConditionalSequentialMover and the NonCanonicalConditionalSequentialMoveChange were derived from existing classes available in OpenPathSampling. The stable states, the dividing surface and other regions are identified via a user-defined reaction coordinate and resulting trajectories are saved in a Storage object.
  • The class ReactiveFluxAnalysis provides functionality to analyze previously generated and stored trajectories similar to its parent class ShootingPointAnalysis. In addition to trajectories the user needs to provide the gradient of the reaction coordinate at the dividing surface. With the stored velocities at the trajectory starting points it is possible to compute the time derivate of the reaction coordinate and therefore (together with results from a free energy calculation) derive the total flux and the flux for each initial snapshot. Methods to visualize e.g. per-snapshot results in 1D- and 2D-histograms are also provided.

Background Information

This module builds on OpenPathSampling, a Python package for path sampling simulations. To learn more about OpenPathSampling, you might be interested in reading:

Testing

To test this module you need to download the source files package (see the Source Code section below) and install it using python setup.py install from the root directory of the package. In the ops_rf/tests folder type nosetests testrfanalysis.py to test the module using the nose package.

Examples

See the rf-example.ipynb IPython notebook in the source directory, here is the direct link: https://gitlab.e-cam2020.eu/Classical-MD_openpathsampling/RF/blob/master/ops_rf/rf-example.ipynb To run the example execute jupyter notebook rf-example.ipynb in your terminal.