Utilities to help with analysing multiple single cell structures. Precomputes distances / depths for each .nuc file. Helper functions for dealing with offset starts etc.
A NucFrame object represents the structure for a single cell Hi-C experiment. It can be created from a .nuc file with the from_nuc method. This will create an hdf5 file with various cached results (for instance, depth). Alternatively, this file can be directly loaded with NucFrame.__init__.
A NucFrames object loads multiple NucFrame objects and ensures they have consistent start/end basepairs. It is created by passing in a list of NucFrame file locations, typically done with glob.glob("/path/to/files/*.hdf5").
Requires python3. Using Anaconda to create a fresh environment, and then manage these requirements, is suggested. All packages can be installed with pip or conda. Anaconda can be downloaded from https://www.continuum.io/downloads.
Having cloned the repository with git clone, run python setup.py install to install. Before running any code, install the dependencies listed
below.
- numpy
- scipy
- numba (>= 0.29.0, installable via
conda install -c https://conda.anaconda.org/numba numba) - h5py
- tqdm
- networkx
- pandas
- hypothesis
Once the package and dependencies are installed, the package can be used. Example analysis scripts using this library can be found here.