The GTEx web portal (http://gtexportal.org) develops specialized web-based visualization tools for exploring tissue-specific gene expression and regulation. Currently, in this public repository, we provide the source code of the following tools:
Also, try the tools at jsFiddle.
Running demos locally on your computer
Prerequisites
- A local copy of this repo.
- Access to the internet, for real-time data retrieval from the GTEx web service and external library dependencies.
- A modern web browser.
Launching the demo
In the repo's root directory on your computer, start up a simple Python HTTP server, and specify a port (e.g. 9090):
python -m SimpleHTTPServer 9090
Once the HTTP server is running, open the demo home page using the following URL in a web browser:
localhost:9090 (or the port of your choice)
Using Rollup to bundle the applications
Most of our tools are written in ES6, with the exception of the GTEx Gene-eQTL Visualizer, and we recommend using a module bundler such as Rollup to recompile code if needed.
Rollup installation
To install Rollup and required libraries, you may run npm install in the repo's root directory on your computer. This will install the libraries under a subdirectory: node_modules.
Rollup configuration
The rollup configuration files for each tool is located in the directory rollup. To recompile a tool (e.g. GTEx Expression Map): run the following command in your local repo's root directory:
rollup -c rollup/rollup.expression-map.config.js
This will recompile and generate a new bundled tool code in the directory build/js/.
To minify the bundled code, first set the environment variable NODE_ENV to "prod", for example in a Bash terminal, the command would be:
export NODE_ENV="prod"
Then run rollup to recompile the code. All the demos are using the minified code of the tools.
GTEx Expression Map
The GTEx Expression Map is an interactive heatmap specifically designed for rendering expression data. The GTEx portal uses this tool to report and summarize multi-gene, multi-tissue expression data.
GTEx Transcript Browser
The GTEx Transcript Browser is for exploring expression of individual transcripts of a gene.
GTEx eQTL Dashboard
The GTEx eQTL Dashboard reports single-tissue eQTLs for a user-provided list of gene-variant pairs and tissues. All eQTL results, including non-significant ones, are visualized in a grid of violin plots grouped horizontally by the gene-variant pairs and aligned vertically by tissues. Each violin plot shows three expression distributions of the genotypes: homozygous reference, heterozygous, and homozygous alternative alleles.
GTEx Gene-eQTL Visualizer
The GTEx Gene-eQTL Visualizer displays single-tissue eQTLs of a gene in a bubble heatmap--the rows are tissues in alphabetical order, and columns are variants sorted by genomic location. The color and size of the bubbles are scaled by the normalized effect size (NES) and p-value of the eQTLs. Only tissues with significant eQTLs associated with the query gene are reported and rendered. The pairwise linkage disequilibrium (LD) values of the variants are also displayed and aligned below the bubble heatmap columns.
Note: The GTEx Gene-eQTL Visualizer is currently in ES5. We will update the code to ES6 in the near future.
Versioning
For the versions available, see the tags on this repository.
GTEx Portal Team
- Katherine Huang
- Kane Hadley
- Duyen Nguyen
- Jared Nedzel
- François Aguet
- Kristin Ardlie
License
This project is licensed under the terms of the BSD 3-clause license - see the LICENSE.md file for details.
Dependencies
GTEx Visualizations is distributed, in part, under and subject to the provisions of licenses for:
D3.js (https://d3js.org/), Copyright (c) 2017 Mike Bostock. All rights reserved. Licensed under the BSD 3-clause license (https://opensource.org/licenses/BSD-3-Clause); and
jQuery (https://jquery.com/), Copyright (c) 2018 The jQuery Foundation. All rights reserved. Licensed under the MIT license (https://jquery.org/license/).
Acknowledgements
The GTEx portal is supported by NIH contract HHSN268201000029C and U41 grant HG009494-01. We also thank our Pfizer collaborators for contributing to visualization discussions and suggestions.