PyHillFit - python code to perform Bayesian inference of Hill curve parameters from dose-response data

Code to load dose-response data and fit dose Hill response curves in a Bayesian inference framework.

This code is associated with the paper "Hierarchical Bayesian inference for ion channel screening dose-response data". Wellcome Open Research 1:6. doi:10.12688/wellcomeopenres.9945.2.

Schematic of inputs and outputs

N.B. The response should be the percentage inhibition.

Pre-requisites

The following python packages are pre-requisites for running PyHillFit:

• numpy
• cma
• matplotlib
• scipy
• pandas
• seaborn

On most linux distributions you can install these via pip, which itself can be installed, if it isn't already present, following the instructions on the pip homepage.

Then all the above dependencies can be installed in one go with:

sudo pip install numpy cma matplotlib scipy pandas seaborn

Crumb dataset

We have made a .csv file of the Crumb et al. dataset, which is available in the data folder, together with some example python scripts for reading it. You can fit your own data by putting them into a similar format to this .csv file. Note that doses/concentrations should be given in microMolar.

Running PyHillFit

To run the python-based dose-response fitting code, see the README in the python folder.

Uncertainty Propagation

To run the Uncertainty Propagation example based on PyHillFit output, see the README in the chaste folder.