This repository contains an ipython notebook script which demonstrates the following:

Calculation of "Void" regions in population Hi-C data - a small proportion of the sequence with unusual numbers of contacts that is excluded from structural analyses.

Calculation of A/B chromosome compartments using population Hi-C data using a simple k-means clustering of the covarience matrix of the observed to expected contact ratio.

Calculation of spatial density enrichment whereby 3D co-localisation of two data tracks on a genome structure is compared to a random/null hypothesis where sequence positions of the data are circularly permuted (thus preserving sequence relationships but sampling diffferent structural conformations). This analysis allows the degree of spatial clustering to be assessed above what is expected from relative sequence positioning of the data, i.e it removes the effect of sequence clustering.

Each of the above is achieved by a separate function within the larger script which also contains all of the necessary utility functions.

After following the installation instructions, the notebook can be run with the command jupyter-notebook in the repository folder.

Dependencies are SciPy, NumPy, matplotlib and Cython. These are all available through pip, and in most distributions' package managers. To view the notebook, a web browser is required.

Jupyter can be installed using pip install jupyter, for either Python3 and Python2.

The notebook is written in Python2, so the Python2 Jupyter kernel must be installed: pip2 install ipykernel followed by python2 -m ipykernel install.

Note: it might be necessary to explicitly select the Python2 kernel from the Kernel menu when running the notebook.

The script requires the demonstration data (available as a .tar.gz archive) to be extracted into the same directory as the notebook script, but otherwise the material is self-contained.