This repository contains scripts for optimising models of the cardiac hERG ion channel to wild type (WT) and R56Q mutant hERG channel electrophysiology data using Myokit and PINTS modules in Python. This code is associated with the paper:
"Electrophysiological characterization of the hERG R56Q LQTS variant and targeted rescue by the activator RPR260243". J Gen Physiol (2021) 153 (10): e202112923. Kemp, J. M., Whittaker, D. G., Venkateshappa, R., Pang, Z., Johal, R., Sergeev, V., Tibbits, G. F., Mirams, G. R., Claydon, T. W.
The instructions in this section assume that you are using a Linux distribution with python3, pip and virutalenv installed. For example, if you are using Ubuntu 20.04.1 LTS, running the command sudo apt update followed by sudo apt install virtualenv python3 will install these packages.
It is recommended to install libraries and run this repository's scripts in a virtual environment to avoid version conflicts between different projects. To create such a virtual environment open a terminal and do the following:
- Clone the repository (for example, by using
git clone https://github.com/CardiacModelling/R56Q-modelling.git) - Type
cd R56Q-modellingto navigate inside the repository - Set up a virtual environment using
virtualenv --python=python3 venv - Activate the virtual environment using
source venv/bin/activate - Install the required packages by typing
pip install -r requirements.txt.
When you are finished working with the repository, type deactivate to exit the virtual environment. The virtual environment can be used again with the activate command as shown above.
For calibration and validation of the mathematical model, the study used a novel collection of protocols (shown below), including the previously-described 'staircase' protocol by (Lei et al., 2019).
The full voltage clamp waveform itself can be found as a comma-separated values (.csv) file in data/Protocols/ called RPR-protocol.csv. This contains a list of voltages (in mV) and times (in ms) that comprise the protocol at a sampling frequency of 10 kHz.
Experimental data obtained from the above electrophysiology protocol (recorded at 37°C in HEK cells) have been processed, separated into constituent protocols, and stored in data/. The data files are named according to [protocol]-[mutant]-cell-[cell number].csv, and are stored in sub-folders according to protocol. The activation protocol data for WT, cell 2 can thus be found in activation-WT-cell-2.csv in the data/Activation/ folder.
The parameter inference in this study uses the CMA-ES algorithm to search the parameter space. To reproduce the parameter fitting, generate fits for WT and R56Q by running
python cmaesfit.pyandpython cmaesfit.py --mutant 2.
Each of these could take several hours running in parallel, so are best performed using a HPC resource. The number of repeats can be reduced from the default 20 using the --repeats input argument, although this may not explore the parameter space sufficiently to find a global minimum. The default model used, described as M10 throughout the code, corresponds to the structure shown below (taken from manuscript Figure 9). The code also supports inference for a C-C-O-I structure hERG model (labelled CCOI in the code), which can be toggled by changing the --model input argument.
Once finished, the results of the CMA-ES optimisation are deposited in cmaesfits/. The maximum likelihood estimate parameters for the default runs can be found in WT-model-M10-fit-staircase1-artefact.txt and R56Q-model-M10-fit-staircase1-artefact.txt for WT and R56Q, respectively. Detailed breakdowns of the scores and parameters for each repeat can be found in the accompanying parameter and log .txt files. For convenience, the results of the inference used for the actual modelling data in the manuscript, along with detailed parameter and score logs, can be found already in cmaesfits/.
Most of the plots found in the manuscript are composite plots consisting of multiple, individually-generated panels. Nonetheless, these can be generated easily using Python scripts provided in figures/. It is recommended to run the scripts with --show to simply check the results on screen, or with --dpi 300 in order to generate high-quality, publication-standard figures in .png format (stored in sub-folders of the figures/ directory). Where results from a single cell are shown, these correspond to cell 2 for the WT recordings, and cell 5 for R56Q (which can be identified based on the filename). In order to quickly generate all panels for a given figure, Bash scripts are provided. To use them, open a terminal in the figures directory and do the following:
- Type
chmod u+x *.shto ensure that the scripts are executable - Manuscript Figure 9A is a schematic drawing. To generate Figure panels 9B-D, run
./figure9-panels.sh - Similarly, to generate panels for Figures 10-13, run
./figure[figure number]-panels.sh
All final, composite figure files (pertaining to the in silico modelling) used in the manuscript are stored in figures/Paper_figures/.
If you publish any work based on the contents of this repository please cite (CITATION file):
Kemp, J. M., Whittaker, D. G., Venkateshappa, R., Pang, Z., Johal, R., Sergeev, V., Tibbits, G. F., Mirams, G. R., Claydon, T. W. (2021). Electrophysiological characterization of the hERG R56Q LQTS variant and targeted rescue by the activator RPR260243. Journal of General Physiology 153 (10): e202112923.
The scientific approach and code structure in this work owe a lot to the following previous, related publications from our lab which may also be of interest:
Beattie, K. A., Hill, A. P., Bardenet, R., Cui, Y., Vandenberg, J. I., Gavaghan, D. J., de Boer, T. P., Mirams, G. R. (2018). Sinusoidal voltage protocols for rapid characterisation of ion channel kinetics. The Journal of Physiology, 596, 1813–1828.
Clerx, M., Beattie, K. A., Gavaghan, D. J., Mirams, G. R. (2019). Four ways to fit an ion channel model. Biophysical Journal, 117, 2420-2437.
Lei, C. L., Clerx, M., Gavaghan, D. J., Polonchuk, L., Mirams, G. R., Wang, K. (2019). Rapid characterisation of hERG channel kinetics I: using an automated high-throughput system. Biophysical Journal, 117, 2438-2454.
Lei, C. L., Clerx, M., Whittaker, D. G., Gavaghan D. J., de Boer, T. P., Mirams, G. R. (2020). Accounting for variability in ion current recordings using a mathematical model of artefacts in voltage-clamp experiments. Philosophical Transactions of the Royal Society A, 378: 20190348.