ravehlab / mmsm

MMSM (Multiscale Markov State Models)

Author:

Kessem Clein (Raveh lab)

Date last updated:

Oct 1st, 2021 or later

Synopsis:

Multiscale Markov-State Models (MMSMs) and MMSM-Explore code

Prerequisties:

• scipy
• numpy
• matplotlib
• sklearn

Folder structure:

base

Abstract base classes defining the interfaces of the different modules

discretizers

Classes that inherit from BaseDiscretizer. These provide methods for discretizing continuous spaces into discrete sets of states.

estimators

This module has two submodules for two different types of estimators used by an HMSM.

metastable_partition

Classes that inherit from MetastablePartition. These provide methods for estimating partitions of the states of a Markov model, such that the resulting partition is a coarse grained estimate of the Markov model.

transitions

Estimators of the transition probabilities between states, based on observed transitions.

models

hmsm

Implementation of the HMSM model, that makes use of the different estimators and optimizers.

parallel_hmsm

Implementation of the HMSM model that makes use of parallelization. (Not implemented)

optimizers

Methods for choosing vertices from an HMSM for adaptive sampling.

plots

Various methods for creating plots.

samplers

Classes that inherit from BaseSampler.

util

Various utility functions used in the package.

Usage:

TBD