D3GB: An Interactive Genome Browser for R,Python,and WordPress

1. Introduction

Genome browsers are essential tools to explore biological information within a genome context. Currently, they are widely used for the visualization of data provided by the generally used applications of next generation sequencing. Some popular genome browsers are, for example, GBrowse (Stein et al., 2002), JBrowse (Buels et al., 2016), Integrative Genomics (Robinson et al., 2011), Apollo (Lee et al., 2013), Artemis (Rutherford et al., 2000), and Trackster (Goecks et al., 2013). Each one has its own interesting features for scientists, such as Web visualization of several data types (i.e., GBrowse and JBrowse), effective visualization of aligned sequencing reads (i.e., IGV), genome annotation and editing (i.e., Apollo and Artemis), and integration with analysis pipelines (i.e., Trackster).

Lower costs and the extensive nature of next generation techniques have led to new challenges for displaying genomes. De novo sequencing of several species or indeed pangenomes is becoming a common practice. These projects require genome browsers with quick and scriptable configurations, which enable the visualization of several genomes without wasting time on loading or sharing. Full integration of genome browsers in analysis pipelines is also a current challenge. This integration allows the optimization of analysis parameters through visualization, guiding the analysis to obtain better results.

Here, we present D3GB (D3 Genome Browser), a new genome browser, for quick and in-browser visualization of genomic data. This software project is available as an R package and a Python module to enable its easy integration in analysis pipelines. It is also distributed as a WordPress plugin, which allows the integration of D3GB in websites with functionalities of WordPress and third-party plugins. D3GB is Web based, requiring only a Web browser to view interactive visualization of genomes. It can be shared with an Apache application server or used in the user local machine without the installation of additional software. (See the first Reference for availability details.)

2. Methods

D3GB has been implemented in JavaScript, R, Python, and PHP programming languages. The front-end is developed with D3, which offers multiple visualization features, vectorial graphs, and an adequate performance for an agile exploration. Two different back-ends were implemented in PHP (for server mode) and JavaScript (for the local mode), which allows the efficient accession to genome sequence and tracks. D3GB uses the following JavaScript libraries: (1) D3.js library (Bostock et al., 2011) for the graphical interface in the front-end, (2) sql.js for the connection between the back-end and the database, and (3) jsPDF for exporting graphics in PDF.

In addition, we developed R, Python, and WordPress setting up interfaces distributed as a package, module, or plugin, respectively. These interfaces generate D3 genome browsers processing popular genomic data formats such us BED, GFF, GenBank, FASTA, or VCF.

3. Results

The present project has fulfilled the initial objective of creating a Web-based interactive genome browser, which facilitates its integration with analysis pipelines. The current version of D3GB implements this integration with Python and R, which are popular programming languages in bioinformatics research. These interfaces allow the configuration of visualization parameters and the generation of genome browsers with the execution of R or Python functions. In this way, analysts can automate the visualization of results and integrate this step in their analysis scripts.

D3GB follows the current developing trends of generating interactive graphs with JavaScript. This technology only needs a current Web browser to view the resulting graphs, and can be shared online with an application server. Figure 1 shows an example of the D3GB graphical interface, and the project website contains some live examples, reference manual, and tutorials. The main features of D3GB are as follows:

(1) Easy integration with analysis protocols and pipelines developed with R or Python.

OPEN IN VIEWER

FIG. 1.

D3GB visual interface. The figure shows a D3GB human genome browser displaying genes, domains, variants with VEP annotation, and expression values. Live example and tutorials for its generation are available on the project website (http://d3gb.usal.es). VEP, variant effect predictor.

D3GB was designed under the idea of easing the automatic generation of genome browsers for computational biologists and simplifying the visualization of genomic data for the final users. From our experience, it is very useful for testing protocols, reporting analysis results, and allowing the exploration of results to nonbioinformatic users. D3GB is a live project that is open to suggestions and external contributions to adapt the project to new scientific requirements.