We would love your help in making MHCflurry a useful resource for the community. No contribution is too small, and we especially appreciate usability improvements like better documentation, tutorials, tests, or code cleanup.
We hope MHCflurry will grow to include reference implementations for state-of-the-art approaches for T cell epitope prediction. This includes pan-allele MHC I and II prediction and closely related tasks such as prediction of antigen processing and immunogenicity. It does not include tasks such as B cell (antibody) epitope prediction, prediction of TCR/pMHC interactions, or downstream tasks such as cancer vaccine design. All committed code to MHCflurry should be suitable for regular research use by practioners. This likely means that new models will require a benchmark evaluation with a publication or preprint before they can be accepted.
If you are contemplating a large contribution, such as the addition of a new predictive model, it probably makes sense to reach out on the Github issue tracker (or email us at [email protected]) to discuss and coordinate the work.
All contributions can be made as pull requests on Github. One of the core developers will review your contribution. As needed the core contributors will also make releases and submit to PyPI.
A few other guidelines:
- Any generated resource, such as trained models, must be associated with a
GENERATE.sh
script in downloads-generation. Running this script with no arguments should fully reproduce the generated result. Reproducability of MHCflurry trained models and related data (such as curated training data, allele sequences, etc.) is key to allowing others to build upon and improve our work. - MHCflurry supports Python 2.7 and 3.3+ on Linux and OS X. We can't guarantee support for Windows. If you are having trouble running MHCflurry on Windows we would appreciate contributions that help us address this.
- All functions should be documented using numpy-style docstrings and associated with unit tests.
- Bugfixes should be accompanied with test that illustrates the bug when feasible.
- Contributions are licensed under Apache 2.0
- Please adhere to our code of conduct.
Working on your first Pull Request? One resource that may be helpful is How to Contribute to an Open Source Project on GitHub.