Variant Effect Prediction for Python
Project description
WARNING: This code is an alpha release and not production-ready. APIs may change at any time.
Veppy is a genetic variant effect predictor for Python. Inspired by SnpEff and VEP.
Installation
$ pip install veppy
Installation from source
$ git clone git@github.com:solvebio/veppy.git
$ cd veppy
$ python setup.py install
Setup
Step 1 (OPTIONAL): Prepare a directory for veppy data
The default data path is: ./data
You can override this by setting $VEPPY_DATA_DIR.
export VEPPY_DATA_DIR=/opt/veppy
Step 2: Download source data and build indexes
NOTE: This step downloads about 1gb of data. After indexing, the data directory will consume about 8gb of disk space.
./scripts/download_data_GRCh37.sh
Step 3: Index the source data
python ./run_index.py
Example Usage
>>> from veppy.veppy import calculate_consequences
>>> variant = ('1', 8025384, 'A', 'T')
>>> result = calculate_consequences('GRCh37', *variant)
>>> print result.results
Testing
Tests are currently based on chr1 versions of input data. Full genome tests are coming soon!
$ nosetests
Coverage:
$ nosetests --with-coverage --cover-package=veppy
About SolveBio
SolveBio is a genomics company based in New York City.