paulsengroup / hictkpy

Python bindings for hictk: read and write .cool and .hic files directly from Python

hictkpy

Python bindings for hictk, a blazing fast toolkit to work with .hic and .cool files.

Installing hictkpy

hictkpy can be installed in various ways. The simples method is using pip: pip install hictkpy.

Refer to Installation for alternative methods.

Using hictkpy

import hictkpy

path_to_clr = "file.mcool"  # "file.hic"

clr = hictkpy.File(path_to_clr, 100_000)
sel = clr.fetch("chr1")

df = sel.to_df()     # Get interactions as a pd.DataFrame
m1 = sel.to_numpy()  # Get interactions as a numpy matrix
m2 = sel.to_coo()    # Get interactions as a scipy.sparse.coo_matrix

For more detailed examples refer to Quickstart.

The complete documentation for hictkpy API is available here.

Citing

If you use hictkpy in you reaserch, please cite the following publication:

Roberto Rossini, Jonas Paulsen hictk: blazing fast toolkit to work with .hic and .cool files. bioRxiv 2023.11.26.568707. https://doi.org/10.1101/2023.11.26.568707

BibTex

@article {hictk,
	author = {Roberto Rossini and Jonas Paulsen},
	title = {hictk: blazing fast toolkit to work with .hic and .cool files},
	elocation-id = {2023.11.26.568707},
	year = {2023},
	doi = {10.1101/2023.11.26.568707},
	publisher = {Cold Spring Harbor Laboratory},
	URL = {https://www.biorxiv.org/content/early/2023/11/27/2023.11.26.568707},
	eprint = {https://www.biorxiv.org/content/early/2023/11/27/2023.11.26.568707.full.pdf},
	journal = {bioRxiv}
}

hictkpy's Issues

PixelSelector should keep alive its file handle

[feature] Implement accessors for balancing weights

Review cooler attribute handling

TODOs:

• Use dash in cooler attribute names
• Ensure missing values are handled properly (e.g. missing bin-type implies bin-type=fixed

[bug] query mirroring is not always correct

Fix mirroring logic for queries like fetch("4:0-987474", "4:0-920561").to_df(), where the start position is the same, but the end position it is not (and the end position of the first set of coordinates is greater than the second one).

This bug was uncovered by the fuzzy tests run as part of hictk's CI, see paulsengroup/hictk#173.