Genome contact map explorer: a platform for the comparison, interactive visualization and analysis of genome contact maps

Abstract

Hi-C experiments generate data in form of large genome contact maps (Hi-C maps). These show that chromosomes are arranged in a hierarchy of three-dimensional compartments. But to understand how these compartments form and by how much they affect genetic processes such as gene regulation, biologists and bioinformaticians need efficient tools to visualize and analyze Hi-C data. However, this is technically challenging because these maps are big. In this paper, we remedied this problem, partly by implementing an efficient file format and developed the genome contact map explorer platform. Apart from tools to process Hi-C data, such as normalization methods and a programmable interface, we made a graphical interface that let users browse, scroll and zoom Hi-C maps to visually search for patterns in the Hi-C data. In the software, it is also possible to browse several maps simultaneously and plot related genomic data. The software is openly accessible to the scientific community.

Hi-C analysis: from data generation to integration

“…They are instead based on the assumption that each genomic locus should have “equal visibility,” i.e., the interaction signal, as measured by Hi-C for each genomic locus, should add up to the same total amount. These include the sequential component normalization (SCN) (Cournac et al 2012 ), the iterative correction and eigenvector decomposition (ICE) normalization (Imakaev et al 2012 ), and Knight-Ruiz matrix-balancing approach (Knight and Ruiz 2013 ; Rao et al 2014 ), implemented by multiple tools (Sauria et al 2015 ; Servant et al 2015 ; Durand et al 2016b ; Kumar et al 2017 ; Wolff et al 2018 ; Stansfield et al 2018 ). ICE normalization has also been optimized for handling large- and high-resolution datasets (Kerpedjiev et al 2018 ).…”

“…Other developed formats are an indexed binary file format called Binary Upper TrianguLar matRix (BUTLR) used for visualization by the 3D Genome Browser (Wang et al 2018b ) and the genome contact map format (“gcmap”) based on HDF5 and used for analysis and visualization by the Genome contact map explorer (Kumar et al 2017 ). Finally, an attempt to create a suite of tools for formats conversion, manipulation, and 2D genomic arithmetic of Hi-C data (similar to bedtools) is pgltools which is based on the paired-genomic-loci data (PGL) format (Greenwald et al 2017 ).…”

“…Several tools address these issues and display Hi-C contact maps as heatmaps, supporting matrices saved in binary formats to allow fast retrieval of the data. Juicebox (Durand et al 2016a ), gcMapExplorer (Kumar et al 2017 ), and HiGlass (Kerpedjiev et al 2018 ) allow to smoothly browse Hi-C heatmaps interactively, to zoom in and out with different resolutions, to visualize maps together with other genomics data such as ChIP-seq and to compare multiple maps in a synchronous way. Juicebox is available both as a desktop and a cloud-based web application named Juicebox.js (Robinson et al 2018 ).…”

HiGlass: Web-based Visual Exploration and Analysis of Genome Interaction Maps

“…Existing tools provide different ways of displaying contact frequencies, such as rectangular heatmaps, triangular heatmaps, arc plots, or circular plots, and different degrees of interactivity ranging from static plotting to interactive zooming and panning, as well as different degrees of integration with other genomic data types [18,[24][25][26][27][28][29]. While tools such as Juicebox [18] and Genome Contact Map Explorer [30] provide synchronized exploration of multiple contact maps, they lack an interface for dynamically arranging the views of several Hi-C datasets, and customizing the levels of synchronization between loci, zoom levels, and samples. Furthermore, none provide an interface for continuous panning and zooming of the sort popularized by web based geographical and road maps.…”

Genomic 3D compartments emerge from unfolding mitotic chromosomes

“…Collecting all contacts, we made an artificial Hi-C map and normalised it with the KR-norm (Knight and Ruiz 2013), as in real Hi-C experiments. Finally, we visualised the artificial Hi-C map in the gcMapExplorer software (Kumar et al 2017) (Fig. 1d).…”