This repository contains the necessary code and raw data to reproduce the analysis and modeling attempts described in our manuscript Joch, Bernal et al. 2026.
The code requires Python libraries commonly used in Cheminformatics as fully indicated in the environment.yml file. For reproducibility purposes, it is recommended to run the code in a dedicated conda environment containing those libraries. The environment can be set up as follows:
$ conda env create -f environment.yml
data_compilation.py is a script to load the data sets and combine them with experimentally determined propensity of Hydride Meisenheimer Complex (HMC) formation.
descriptors.py contains a custom Python class to process the descriptors set generated by Jaguar and compiled by Schrödinger's Python script qm_descriptors.py.
utils.py is a module containing several utilities for feature selection, common statitical tests for multiple comparison, and plotting.
The notebooks folder contains Jupyter notebooks showing the descriptive analysis and the predictive modeling attempts presented in the paper.
@article{,
doi = {10.1016/j.bioorg.2026.110172},
url = {https://www.sciencedirect.com/science/article/pii/S004520682600708X},
year = {2026},
month = {September},
publisher = {Elsevier},
volume = {180},
pages = {110172},
author = {Melanie Joch, Freddy A. Bernal, Héctor Torres-Gómez, Valerie Kerndl, Florian Kloss},
title = {Hydride-Meisenheimer metabolism in mammals - scope of an overlooked biotransformation pathway},
journal = {Bioorganic Chemistry}
}