vrs
vrs copied to clipboard
Extensible specification for representing and uniquely identifying biological sequence variation
The GA4GH Variation Representation Specification provides a comprehensive framework for the computational representation of biological sequence variation. VRS is the result of a collaboration among contributors representing national information resource providers, major international public initiatives, and diagnostic testing laboratories.
VRS is licensed under the Apache License 2.0.
NOTE: VRS is under active development. See VR Project Roadmap.
Specific goals:
- Develop common language- and protocol-neutral information models and nomenclature for biological sequence variation.
- From the information models, develop data schemas. The current schema is defined in JSON Schema, but other formats are expected.
- Provide algorithmic guidance and conventions to minimize representational ambiguity.
- Define a globally unique computed identifier for covered data classes.
- Develop validation tests to ensure consistency of implementations.
The VRS model is the product of the GA4GH Variation Representation group.
SEE ALSO: See vrs-python for an implementation and Jupyter notebooks.
Using the schema
The schema is available in the schema/
directory, in both yaml and
json versions. It conforms to JSON Schema draft-07. For a list of
libraries that support JSON schema, see JSON
Schema>Implementations.
Contributing to the schema
VRS uses yaml as the source document for JSON Schema
To convert vrs.yaml to vrs.json:
make vrs.json
You'll probably have to pip install pyyaml
first.
To watch for changes and update automatically:
make -C schema watch &
Contributing docs
The VR specification documentation is written in reStructuredText and
located in docs/source/
. Commits to this repo are built
automatically at vrs.ga4gh.org
.
To build documentation locally, type:
make -C docs clean watch &
Then, open docs/build/html/index.html
. The above make command
should build docs when source changes. (Some types of changes require
recleaning and building.)
Testing
The VRS repo contains two kinds of tests. Basic smoketests in tests/
ensure that the schema is parseable and works with certain tools.
These tests provide a basic sanity check during development.
Validation tests (in validation/
) provide language-neutral tests for
those implementing tools with VRS.
Using smoketests
The smoketests require python 3.8+. This is the recommended setup:
$ python3 -m venv venv
$ source venv/bin/activate
$ pip install -U setuptools pip
$ pip install -r .requirements.txt
$ python3 -m pytest