Computational tools for the multiscale analysis of Hi-C data in bacterial chromosomes

TL;DR

This paper introduces a multiscale analysis framework for bacterial Hi-C data, combining standard and novel tools, with accessible code and statistical assessment, demonstrated on Pseudomonas aeruginosa.

Contribution

It provides a comprehensive Python-based toolkit for multiscale bacterial chromosome analysis, including novel indices and statistical validation methods.

Findings

Identification of nested interaction domains in bacterial chromosomes

Development of a scale-independent locus involvement index

Accessible tools and statistical methods for Hi-C data analysis

Abstract

Just as in eukaryotes, high-throughput chromosome conformation capture (Hi-C) data have revealed nested organizations of bacterial chromosomes into overlapping interaction domains. In this chapter, we present a multiscale analysis framework aiming at capturing and quantifying these properties. These include both standard tools (e.g. contact laws) and novel ones such as an index that allows identifying loci involved in domain formation independently of the structuring scale at play. Our objective is two-fold. On the one hand, we aim at providing a full, understandable Python/Jupyter-based code which can be used by both computer scientists as well as biologists with no advanced computational background. On the other hand, we discuss statistical issues inherent to Hi-C data analysis, focusing more particularly on how to properly assess the statistical significance of results. As a pedagogical example, we analyze data produced in {\it Pseudomonas aeruginosa}, a model pathogenetic bacterium. All files (codes and input data) can be found on a github repository. We have also embedded the files into a Binder package so that the full analysis can be run on any machine through internet.