Automatic detection of equilibrated regions of molecular simulations

A preprint manuscript of this work is available on bioRxiv.

• examples/ - code to generate simulation data and figures used in the manuscript
• manuscript/ - the manuscript source and figures
• references/ - some references cited in the paper

• Check out a copy of this repository, or download a release corresponding to the version of the paper of interest.
• Go to examples/liquid-argon/ and check the README.md for instructions on running the ./reproduce.sh script.

To detect equilibration start times in your own code, you can use the detectEquilibration() method in the timeseries module of pymbar.

You can install pymbar via the conda package manager:

conda install -c omnia pymbar

and then use the detectEquilibration() method:

from pymbar import timeseries
[t0, g, Neff_max] = timeseries.detectEquilibration(A_t)
A_t_equlibrated = A_t[t0:]

where A_t is the observable timeseries and t_0 is the beginning of the equilibrated region.

The equilibration.py module provided in this repo also provides a stand-alone implementation of this feature with a broader choice of methods for computing integrated autocorrelation times.