Michael Clerx, Pieter Collins, Enno de Lange, and Paul G.A. Volders (2016). Myokit: A simple interface to cardiac cellular electrophysiology. Progress in Biophysics and Molecular Biology, Volume 120, issues 1-3, pages 100-114. 10.1016/j.pbiomolbio.2015.12.008.
This repository contains up-to-date example code, adapted from the files originally published with the paper.
An updated version of the schematic describing Myokit is shown below.
All examples can be viewed on github, without downloading. However, if you want to clone this repository, remember the models submodule and use:
git clone https://github.com/myokit/pbmb-2016.git --recursive
Called example-ttp-1-model.mmt in the paper.
The first example can be opened in the Myokit IDE (the Myokit "app") and contains a model, protocol, and script: example-1a-ide.mmt. To use it, download the file by right clicking the link above and selecting "Save link as", then open it in the IDE.
When writing scripts inside an mmt file, like in the example above, you can access the model and protocol parts with:
m = get_model()
p = get_protocol()
In many cases it's better to write a script separately from the mmt file (for example to re-use the exact same model in multiple simulations).
To access an mmt file's model and protocol externally, use:
m = myokit.load_model('my-model.mmt')
p = myokit.load_protocol('my-protocol.mmt')
or
m, p, _ = myokit.load('my-model.mmt')
if both are in the same file (this also loads the script part as _).
A version of example 1 that loads the model and protocol externally is given in the jupyter notebook example-1b-notebook.ipynb. You can download this and the required model file to try offline, or you can view the example in your browser.
Called example-ttp-2-transmural-differences.mmt in the paper.
The second example follows up from example 1b, and shows how to change a model variable in a simulation. Some models, like the Ten Tusscher et al. one used in this example, have a variable used as a "mode switch". Here, we use this to show the three different cell types the model can represent.
The example also shows how to save simulation data to CSV files and load it again for later processing.
Called example-ttp-3-gto.mmt in the paper.
The third example uses the same technique as the second, but this time to explore the effects of varying the Ito conductance in the epicardial model. It also shows how to modify a model after loading it.
Called example-ttp-4-sensitivity.mmt in the paper.
The fourth example does something more mathematical, and shows how Myokit can be used to calculate the partial derivatives of dependent variables (e.g. state variables or currents) on independent variables (e.g. conductance parameters).
Called example-ttp-5-transmural-baseline.mmt in the paper.
In example 5 a strand consisting of 60 endocardial, 45 mid-myocardial, and 60 epicardial cells is simulated, with pacing from the endocardial end.
The example can be viewed in any browser, but re-running the simulations requires a working OpenCL installation
Called example-ttp-6-transmural-modified in the paper.
The sixth example follows up from example 5, but adds 50% ICaL block 240% Ito increase, and a reduced cell-to-cell conductance. This results in a gradual loss of the spike-and-dome shape of the APs.
Called example-ttp-7-transmural-plane.mmt in the paper.
In example 7 we go one step further and simulate a 2d slice of tissue.
Called example-multi-model-testing.mmt in the paper.
Example 8 shows how to use multiple models, without having (a lot of) hardcoded information about each model.
Called example-parameter-estimation.mmt and example-parameter-estimation-2.mmt in the paper.
In 2016 Myokit contained methods for parameter estimation. These have since been merged into the PINTS library.
Examples of using PINTS and Myokit to perform parameter estimation in ion current models are provided in a seperate repository.