Skip to content

Latest commit

 

History

History
64 lines (34 loc) · 3.49 KB

README.md

File metadata and controls

64 lines (34 loc) · 3.49 KB

BioGraph

This is the codebase for the BioGraph genomic analysis platform. It is released under the BSD 2-clause license. See the LICENSE file for details.

Quick start

Download a pre-built binary from GitHub releases. Releases are statically built and should run on most recent Linux distributions.

In addition to the release, you will need a copy of the classifier model available on Zenodo

You will also need a few common bioinformatics packages. To install them on Ubuntu:

sudo apt install -y vcftools tabix bcftools

BioGraph can also be run as a Docker container:

docker run spiralgenetics/biograph ...

Documentation

Full documentation on the GitHub wiki

For other dependencies and full installation instructions, see the online installation docs.

Building the code

We recommend building on Ubuntu 18.04 or later. The following system dependencies are required:

sudo apt-get install -y python3.6 python3.6-dev python3-distutils python3-apt python3-virtualenv virtualenv build-essential libbz2-dev libz-dev libcurl4-openssl-dev dh-autoreconf wamerican docker.io 

Note: Ubuntu 18.04.5 and later no longer include the python command by default. It is safest to avoid the system python and build from inside a virtualenv instead:

virtualenv --python=python3.6 ~/buildenv
. ~/buildenv/bin/activate

You will also need a recent version of bazelisk. Download the binary for your architecture and save it to a file named bazel somewhere in your PATH.

Build and run all tests:

bazel test ... --config=manylinux2014

Create a release tarball:

bazel build -c opt --config=release //python/biograph:package

The pip installable tarball will then be inside bazel-bin/python/biograph/BioGraph*gz

Additional Resources

BioGraph Primer A primer on the challenges in population scale bioinformatics and the design choices behind BioGraph.

White Paper Publication 2020 Leveraging a WGS compression and indexing format with dynamic graph references to call structural variants.

ASHG 2019 A graph genome of the Arab population with base pair accurate structural variation for population genotyping. A collaboration with SIDRA Medical.

AGBT 2017 BioGraph: a reference-agnostic and rapidly queryable NGS read data format enabling flexible analysis at scale. A collaboration with University of Texas Health Science Center and Baylor College of Medicine

HUGO 2016 Using read overlap assembly to accurately identify structural genetic differences in an Ashkenazi Jewish trio. A collaboration with the Baylor College of Medicine Human Genome Sequencing Center.

AGBT 2016 BioGraph Suite: A format to analyze multiple graph genomes created directly from read data