GeneNetwork

This repository contains the current source code for GeneNetwork (GN) (https://www.genenetwork.org/ (version 2). GN2 is a Web 2.0-style framework that includes data and computational tools for online genetics and genomic analysis of many different populations and many types of molecular, cellular, and physiological data. The system is used by scientists and clinicians in the field of precision health care and systems genetics. GN and its predecessors have been in operation since Jan 1994, making it one of the longest-lived web services in biomedical research (https://en.wikipedia.org/wiki/GeneNetwork, and see a partial list of publications using GN and its predecessor, WebQTL (https://genenetwork.org/references/).

Run

We recommend you use GNU Guix. GNU Guix allows you to deploy GeneNetwork2 and dependencies as a self contained unit on any machine. The database can be run separately as well as the source tree (for developers).

Make sure you have the guix-bioinformatics channel set up. Then, to drop into a development environment with all dependencies, run

guix shell -Df guix.scm

Or, to drop into a development environment in a container, run

guix shell -C --network -Df guix.scm

In the development environment, start GeneNetwork2 by running, for example,

env SERVER_PORT=5300 \
    GENENETWORK_FILES=~/data/gn2_data/ \
    GN_PROXY_URL="http://localhost:8080"\
    GN3_LOCAL_URL="http://localhost:8081"\
    ./bin/genenetwork2 ./etc/default_settings.py -gunicorn-dev

For full examples (you may need to set a number of environment variables), including running scripts and a Python REPL, also see the startup script ./bin/genenetwork2.

Also mariadb and redis need to be running, see INSTALL.

Testing

To have tests pass, the redis and mariadb instance should be running, because of asserts sprinkled in the code base.

Right now, the only tests running in CI are unittests. Please make sure the existing unittests are green when submitting a PR.

See ./bin/genenetwork2 for more details.

Mechanical Rob

We are building 'Mechanical Rob' automated testing using Python requests which can be run with:

env ./bin/genenetwork2 \
    GN_PROXY_URL="http://localhost:8080" \
    GN3_LOCAL_URL="http://localhost:8081 "\
    ./etc/default_settings.py -c \
    ../test/requests/test-website.py -a http://localhost:5003

The ./bin/genenetwork2 script sets up the environment and executes test-website.py in a Python interpreter. The -a switch says to run all tests and the URL points to the running GN2 http server.

Unit tests

To run unittests, first cd into the genenetwork2 directory:

# You can use the coverage tool to run the tests
# You could omit the -v which makes the output verbose
runcmd coverage run -m unittest discover -v

# Alternatively, you could run the unittests using:
runpython -m unittest discover -v

# To generate a report in wqflask/coverage_html_report/:
runcmd coverage html

The runcmd and runpython are shell aliases defined in the following way:

alias runpython="env TMPDIR=/tmp SERVER_PORT=5004 GENENETWORK_FILES=/gnu/data/gn2_data/ GN_PROXY_URL="http://localhost:8080" GN3_LOCAL_URL="http://localhost:8081" ./bin/genenetwork2

alias runcmd="time env TMPDIR=//tmp SERVER_PORT=5004 GENENETWORK_FILES=/gnu/data/gn2_data/ GN_PROXY_URL="http://localhost:8080" GN3_LOCAL_URL="http://localhost:8081" ./bin/genenetwork2 ./etc/default_settings.py -cli"

Replace some of the env variables as per your use case.

Troubleshooting

If the menu does not pop up check your GN2_BASE_URL. E.g.

curl http://gn2-pjotr.genenetwork.org/api/v_pre1/gen_dropdown

check the logs. If there is ERROR 1054 (42S22): Unknown column 'InbredSet.Family' in 'field list' it may be you are trying the small database.

Documentation

User documentation can be found here. The architecture of the software stack is described here. The database schema is (still) shared with GN1 and currently described here. Software documentation is being expanded in the source code repository.

Contributing

Issues can be raised through github.

Contribute to GN2 source code by forking the github repository with git and sending us pull requests.

For development GN2 has a mailing list and an active IRC channel #genenetwork on freenode.net with a web interface.

License

The GeneNetwork2 source code is released under the Affero General Public License 3 (AGPLv3). See LICENSE.txt.

More information

For more information visit http://www.genenetwork.org/

Cite

GeneNetwork was published in the Journal of Open Source Software as 'GeneNetwork: framework for web-based genetics' by Zachary Sloan, Danny Arends, Karl W. Broman, Arthur Centeno, Nicholas Furlotte, Harm Nijveen, Lei Yan, Xiang Zhou, Robert W. WIlliams and Pjotr Prins

You may also cite the software using

.

Contact

IRC on #genenetwork on irc.freenode.net.

Code and primary web service managed by Dr. Robert W. Williams and the University of Tennessee Health Science Center, Memphis TN, USA.