smrfeld / physics-based-ml-reaction-networks

Code for paper "Physics-based machine learning for modeling IP3 induced calcium oscillations" - DOI: 10.5281/zenodo.4839127

Physics-based machine learning for modeling IP3 induced calcium oscillations

This repo. contains the code for the paper Physics-based machine learning for modeling IP3 induced calcium oscillations.

Contents

• learn_ip3 - Models extrapolating in IP3 for a single volume and number of IP3Rs (IP3 receptors).
• learn_ip3r - Models extrapolating in the number of IP3Rs for a single volume, across all IP3 values.
• stochastic_simulations - Stochastic simulations for generating training data for both models.

The code is a mix of Mathematica notebooks and C++.

Data

All data is stored using DVC (data version control). You need to install DVC, then you should be able to just run from this directory:

dvc pull

to get all the data.

BE AWARE that this can be a lot to download, particularly because of the stochastic_simulations - otherwise you can look for the .dvc files in the various directory and load those individually.

Figures