Cardiomyocyte Emulator Training Data

The data presented in this repository is the training data (data set #1) used in our work Neural network emulation of the human ventricular cardiomyocyte action potential: A tool for more efficient computation in pharmacological studies, available at https://doi.org/10.7554/eLife.91911. The data set was generated by computing 40,000 cardiomyocyte simulations using the ToR-ORd ionic model by Tomek et al. (ToR-ORd-dynCl: an update of the ToR-ORd model of human ventricular cardiomyocyte with dynamic intracellular chloride). It contains both the used maximum conductances and corresponding action potentials (APs).

Usage

The data was created to be easily read into python or matlab based scripts. The file aps.npz and aps.mat contain the two arrays t of shape [T] and aps of shape [N, T]. A python demo script to load the data is given below, and another one for matlab can be found in the README.

Python

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

#Loading the data
max_conds_df = pd.read_csv("max_conds.csv").drop("db_index", axis=1) #Ignore the db index
with np.load("aps.npz") as f:
    t, aps = f["t"], f["aps"]

#Selecting a random AP
nr_aps = max_conds_df.shape[0]
ap_i = np.random.choice(nr_aps)

#Plotting/printing the random AP and its maximum conductances
print("Visualizing the simulated AP for the maximum conductances:")
print(max_conds_df.iloc[ap_i])
plt.figure()
plt.plot(t, aps[ap_i])
plt.show()

Citation

If you use these data in your research, please cite the publication from which the data originated:

Grandits, T., Augustin, C. M., Haase, G., Jost, N., Mirams, G. R., Niederer, S. A., Plank, G., Varró, A., Virág, L., & Jung, A. (2023). Neural network emulation of the human ventricular cardiomyocyte action potential: A tool for more efficient computation in pharmacological studies. eLife, 12. https://doi.org/10.7554/eLife.91911

@article{grandits_neural_2023,
    title = {Neural network emulation of the human ventricular cardiomyocyte action potential: a tool for more efficient computation in pharmacological studies},
    volume = {12},
    shorttitle = {Neural network emulation of the human ventricular cardiomyocyte action potential},
    url = {https://elifesciences.org/reviewed-preprints/91911},
    doi = {10.7554/eLife.91911},
    language = {en},
    journal = {eLife},
    author = {Grandits, Thomas and Augustin, Christoph M. and Haase, Gundolf and Jost, Norbert and Mirams, Gary R. and Niederer, Steven A. and Plank, Gernot and Varró, András and Virág, László and Jung, Alexander},
    month = dec,
    year = {2023}
}