Hi-C Notes

Cooler wraps around HDF5 files (use h5py package for python)
.mcool adds an additional layer of organization to combine multiple cooler files into one
- This is like an h5 file. Just use the .keys() command to identify what additional layer(s) there are.
- At the bottom layer, a single cooler file should have ['bins', 'chroms', 'indexes', 'pixels'] as the keys
Cooler file is organized like a dictionary
Useful commands:
- c=cooler.Cooler(filepath) creates an cooler object. This object is serialized and is only opened when needed. c will be referred as the general cooler file
- c.info prints out everything about this cooler file
- c.chromnames prints out the chromosome names
- c.chromsizes prints out the sizes of each chromosome
- c.bins() access the bins. This is sliceable like a datafra me. For instance, c.bins()[:5]
- c.bins()['weight'] returns the weight for each bin
- c.pixels() returns the non-zero upper triangle entries. If using c.pixels(join=True), then the bins IDs are expanded into genomic bin coordinates. The object returned by this command is a pandas dataframe, easy to dump into file.
- c.matrix() returns a 2D numpy array. c.matrix(balance=True) will normalize the matrix.
- To visualize the contact matrix, we can use the following command
  - ```
  fig = plt.figure(figsize=(10, 10))
  ax = fig.add_subplot(111)
  im = ax.matshow(np.log10(arr), cmap='YlOrRd')
  fig.colorbar(im)
```

chuckzzzz/HiC