Genetics API Prototype

[hammer]

(Posted by @eric-czech)

We've been working on prototyping some ideas for a PyData-based API covering a broad range of use cases in genetics. This focuses on the details of the frontend for an API intended to complement the bottom-up proposal by @alimanfoo in skallel. Feedback would be helpful for understanding applications implicitly excluded here or implementation details overlooked, and ideally this topic results in a consensus on what parts of this are missing, worth iterating on, and/or merging into skallel.

Links to notebooks and code for this are here: Xarray Genetic API Prototype, and further details/considerations underlying this approach are being tracked in this issue.

The related post on Design principles for a new GWAS Toolkit is a better place for higher level design questions, but this is a good thread for noting specific limitations of Xarray, problems with dispatching to array backends, how to enforce conventions on naming and indexing, whether or not this IO extension mechanism is too simplistic, etc.

[hammer]

(Posted by @alimanfoo)

Hi @eczech, just scanned through the notebooks in the API prototype, quick message to say this is so interesting, thanks for posting. I'll work through these in more detail and hope to follow up.

[hammer]

(Posted by @eric-czech)

Sure thing! Let me know if you have any concerns/questions.