StripePy: fast and robust characterization of architectural stripes

TL;DR

StripePy is a new tool for identifying architectural stripes in genomic data, offering faster and more reliable analysis than existing methods.

Contribution

StripePy introduces a computational geometry-based method for stripe detection and includes a new simulated benchmark called StripeBench.

Findings

StripePy outperforms existing tools in detecting architectural stripes in Hi-C and Micro-C data.

StripePy includes a simulated benchmark called StripeBench for testing and validation.

Abstract

Architectural stripes in Hi-C and related data are crucial for gene regulation, development, and DNA repair. Despite their importance, few tools exist for automatic stripe detection.

We introduce StripePy, which leverages computational geometry methods to identify and analyze architectural stripes in contact maps from Chromosome Conformation Capture experiments like Hi-C and Micro-C. StripePy outperforms existing tools, as shown through tests on various datasets and a newly developed simulated benchmark, StripeBench, providing a valuable resource for the community.

StripePy is released to the public as an open-source, MIT-licensed Python application. StripePy source code is hosted on GitHub at https://github.com/paulsengroup/StripePy and is archived on Zenodo. StripePy can be easily installed from source or PyPI using pip and from Bioconda using conda. Containerized versions of StripePy are regularly published on DockerHub.

Figures