Schema and API for a relational db that encodes gene models in an explicit, structured, and robust fashion.
GeenuFF is currently extremely beta and very unstable. We're keen to get feed back or ideas from the community (even if it's just whether you think this could be useful to you if developed further), but if you build on GeenuFF as it is now, you're doing so at your own risk.
We developed this to provide a way of unambiguously encoding gene models, (the way the DNA sequence is interpreted to produce proteins) that is both robust to partial information and biological complexity.
A more extensive description can be found in
https://weberlab-hhu.github.io/GeenuFF/
GeenuFF has been tested exclusively (so far) in python3.6
I would recommend installation in a virtual envinronment. https://docs.python-guide.org/dev/virtualenvs/
The dustdas dependency is available from github. From a directory of your choice (and preferably in a virtualenv):
git clone https://github.com/janinamass/dustdas.git
cd dustdas
python setup.py install
cd ..Afterwards install GeenuFF in the same fashion.
git clone https://github.com/weberlab-hhu/GeenuFF.git
cd GeenuFF
pip install -r requirements.txt
python setup.py install
cd ..And you might want to run the tests (sorry for the strict directory, will fix)
cd GeenuFF/geenuff
py.test
cd ../..- Add a validation module to check structure of gene models.
- Add extraction of raw & mature transcript, CDS, and protein sequence as a demo application.
- Visualization.
To @janinamass for discussion and advice.