Introduction

The adaptive_sampling package

This page provides an introduction to the adaptive_sampling package, which is designed to facilitate reaction space exploration and importance sampling in molecular dynamics simulations. The package includes the following subpackages:

• sampling_tools: Contains tools for importance sampling of molecular transitions, facilitating the calculation of associated reaction and activation free energies.

• exploration_tools: Provides tools for exploring the chemical space of molecular systems using the ab initio Nanoreactor or Hyperreactor dynamics.

• processing_tools: Provides tools for processing and analyzing simulation data, including methods for the automatic generation of reaction networks from exploration runs and methods to calculate unbiased ensemble properties from importance sampling runs.

• interface: Contains interfaces for integrating with molecular dynamics engines. Available are interfaces to OpenMM (for MM simulations), ASE (for QM simulations) and ASH (for QM/MM simulations).

• colvars: Provides a large set of collective variable (CVs) for the sampling tools, including common CVs like distance, angle, and dihedrals, and more complex ones like path CVs and machine learning CVs.

The package is developed in the research group of Prof. Dr. Christian Ochsenfeld at LMU Munich.

Warning

The adaptive_sampling package is under active development. Please report any issues or bugs you encounter on GitHub.