ChromoMap: an R package for interactive visualization of multi-omics data and annotation of chromosomes

1Recent dramatic decrease in Next Generation Sequencing costs, and its application to the analysis of multiple levels of biological information (genomics, transcriptomics, epigenomics), has rapidly increased the number of multi-omics datasets available in public repositories. This has sifted the interest of researchers towards the annotation and comparison of such information at a genome or chromosome level. Besides identifying structural differences, comparison of genomic annotations can help understanding how variability within and between species affect phenotypes.

2Currently, genomes are viewed using interactive web-based genome browser applications like JBrowse. There is a paucity of non-web based independent tools capable of generating publication-ready interactive visualization of annotated genomes. This is more evident with the recent advent of diploid phased genomes of highly heterozygous samples. The R package chromoMap, is capable of annotating multiple genomic features, for the generation of interactive graphics. This allows the visualization in a single plot of chromosomes or chromosome sections of any living organism. The large genome size of certain species presents practical graphical challenges of displaying whole chromosomes within the available canvas space suitable for publications. ChromoMap, renders chromosomes as a continuous composition of windows, to surmount this restriction. Each window, consisting of a specific genomic range determined algorithmically based on chromosome length, displays information about the annotation in that region as a tooltip, allowing for plot interactivity.

3chromoMap is built using two languages, R and JavaScript. The R code, written using a procedural approach (as opposed to an object-oriented paradigm), handles the data input and data pre-processing part. The graphics rendering is entirely handled by the JavaScript part, constructed using vanilla JavaScript and D3 JavaScript library (version 4.0). The package was designed and tested using R version 4.0 and is compatible to run with R version 4.0 (or later) on multiple OS (Windows/macOS/Linux). R was chosen due its open-source nature as well as due to its general preference within the bioinformatics community. Annotated chromosomes are constructed by providing genomic co-ordinates, and optional secondary data, as either tab-delimited BED files or directly through R objects. A single R function is used for constructing a variety of different plots by changing its arguments values. For example, Fig. 1 was generated using the code below.

4As shown in the example code above, the user inputs to the function fall into two categories: (1) the BED files or R objects, which specify the genomic coordinates of the chromosomes/chromosomal regions as well as the genomic coordinates for the studied genomic features, (genes, SNPs etc.), along with their associated data (e.g. gene expression values), and (2) the various ‘option’ arguments that are used to specify the graphic-properties of the plot, used for instance to turn a feature ON/OFF etc. Errors caused by the first type, such as user-provided files with incorrect structure or files created in an unsupported format, is handled in the implementation such that the program will first validate the file and will terminate the program with message on the console if the input file was incorrect. In addition to the detection of unsupported file format and/or structure, we implemented an out-of-bound annotations (i.e., annotation coordinates outside the boundary of the target chromosome or genomic feature) detection step. First, out-of-bound annotations are identified by the program and removed from the analysis. Then, the plot is rendered with the remaining annotations, together with an error message, displayed on the console to alert the user of the presence of such errors in their input files. The errors caused by the second type of inputs (e.g. the use of options not available for a given plot type, or failure to specify an option), are also handled in the implementation in a similar way (i.e., checking and exiting the program if encounter errors).

5The generated plot, as viewed in RStudio’s viewer pane, offers users the possibility to zoom into a locus allowing them to explore their annotations of interest. Users can export the plot either as a static image, or as a stand-alone HTML file (see example in https://pouya-dini.github.io/equine-gene-db/#tab2) that contains an interactive plot that can be included in webpage accessible formats (e.g. Shiny Apps), as additional files in publications (See example Additional files 1 and 2), or embedded in RMarkdown documents.

6ChromoMap provides the choice of two annotation algorithms, point-annotation and segment-annotation, differing in how annotations are visualized on the plot. Point-annotation ignores the element’s size, annotating it on a single base. Segment-annotation uses the element’s size to delimit its location. This can be advantageous when visualizing and annotating chromosomal regions or structural elements (Fig. 2).

7Visualizing annotations might require depicting annotations of groups of elements as opposed to individual ones. Groups can be color-coded for effective visualizations (Fig. 3).

8Allow the user viewing information embedded in tooltips associated to genomic features and to be redirected to preexisting webpages containing detailed information about such features. Additionally, users can use the ‘labelling’ option to display labels on static visualizations.

9Feature-associated numeric data, such as gene expression, methylation status, copy number variants, feature density values, etc., can either be summarized as scatter/bar plots (Fig. 4a) or visualized as heatmaps (Fig. 4b). As each window represents a specific range of base pairs, multiple elements can be annotated within its range. ChromoMap uses aggregated data values (sum, average, min, max or count) for each window encompassing more than one element. Individual data values, for each element, can also be viewed in tooltips. Additionally, there is an ‘epi-tag’ feature that allows condition-based marking of loci. More advanced features include creating multidimensional scatter plots (Fig. 1-track 5) or applying mathematical filters on scatter/bar plots as depicted on Fig. 3 by Chidzanga et al..

10This feature allows rendering each chromosome set independently irrespectively of the species’ ploidy. This allows plotting chromosome sets that differ in size and number. This feature also allows the visualization of chromosome sets of different species for comparative genomic studies; or the visualization of highly heterozygous diploid genomes and homologous chromosomes pairs in phased diploid/polyploid genome assemblies (Fig. 1). This feature allows the visualization of biologically important variability that would be lost in a consensus sequence assembly.

11The linkage/correlation between any two annotated features can be visualized either as directed or undirected edges or as thick chord-like connectors. Moreover, data associated to these links can be visualized as color-coded links along with respective legends (Fig. 5). This feature can have several applications like for visualizing the homolog genes or to visualize co-expressed genes.

12ChromoMap is an efficient and user-friendly tool to visualise genomic elements, as well as its associated data (such as multi-omics data), in relation to their regional occurrence across chromosomes for any living organism with an available genome assembly. ChromoMap’s flexible plots permits the comparison of data uniquely associated to each strand of a given chromosome, and of homologous chromosome independently of their size. ChromoMap will allow the individual visualization of multi-omics data in all homologous chromosomes of phased diploid/polyploid genomes.

13ChromoMap is available under the GPL-3 Open Source license from: https://CRAN.R-project.org/package=chromoMap. A vignette is included for comprehensive description of its features. ChromoMap is also available as a containerized web application, referred as chromoMap App, that provides a GUI that will allow researchers to utilize all the above-mentioned features in an R-independent manner. The app can be run locally on the user’s computer. chromoMap App is available as a Docker Image at Docker Hub (https://hub.docker.com/r/lakshay57/chromomap-app). The datasets used and/or analyzed during the current study are available from the corresponding references.

14Project name: chromoMap.

15Project home page: https://lakshay-anand.github.io/chromoMap/index.html.

16Operating system(s): Platform independent.

17Programming language: R, JavaScript.

18Other requirements: None.

19License: GPL3.

20Any restrictions to use by non-academics: license needed.