Introduction

MOXελ is a Python package to ease and accelerate through parallelization, the calculation of voxelized potential energy surfaces, or simply energy voxels, with emphasis on reticular chemistry.

These energy voxels can then be used as a raw input for Machine Learning (ML) algorithms or for manual feature extraction.

Why ΜΟΧελ?

First of all, why energy voxels?

Interactions are the quintessence of chemistry, completely characterizing the behavior (properties) of a material. The potential energy surface (PES) succinctly describes these interactions and as such, it is the most informative input that can be fed into a ML algorithm. Since the PES is a continuous function, voxelization is necessary to convert the PES into a machine understandable format.

Remember, there is no free lunch! Getting the most informative input is computationally expensive. Fortunately, since all calculations in MOXελ are parallelized, the computational cost is significantly decreased.

All you need is a .cif !

Objective

The majority of time in a ML workflow goes into constructing the inputs and making sure they are clean, rather than focusing on the ML part itself.

MOXελ aims to provide a simple and fast interface to generate energy voxels in a ML-ready format, minimizing as much as possible the time spent on these preprocessing steps.

Please note that MOXελ focuses only on energy voxels. If you want to fill the voxels with other values or employ a different featurization scheme, check mofdscribe.

Citing MOXελ

If you use ΜΟΧελ in your research, please consider citing the following work:

@article{Sarikas2024,
title = {Gas adsorption meets deep learning: voxelizing the potential energy surface of metal-organic frameworks},
volume = {14},
ISSN = {2045-2322},
url = {http://dx.doi.org/10.1038/s41598-023-50309-8},
DOI = {10.1038/s41598-023-50309-8},
number = {1},
journal = {Scientific Reports},
publisher = {Springer Science and Business Media LLC},
author = {Sarikas,  Antonios P. and Gkagkas,  Konstantinos and Froudakis,  George E.},
year = {2024},
month = jan
}

TODO

Improve performance
Improve voxelization scheme
Improve modeling of interactions

License

MOXελ is released under the GNU General Public License v3.0 only.

Contents: