Hackathon: add `generic_ir_from_biocypher()` function to ingest TCR data

[slobentanzer]

We will prepare a PR that generalises datasets.vdjdb() to ingest any immune receptor data that was preprocessed by BioCypher, to yield a generic AnnData object that can be flexibly updated and maintained.

Name of the function up for debate. @dbdimitrov

[grst]

Hi @slobentanzer,

has anything happend in the meanwhile on integrating BCR/TCR data into Biocypher? I am asking this because in #469 it surfaces that providing access to reference databases becomes a real data curation effort rather than just downloading and transforming a table. It would be nice to separate this from scirpy and have scirpy just obtain this from a knowledge graph.

This would also require that there's an API to a publicly available biocypher knowledge graph - I'm not even sure if this is a plan at the moment?

[slobentanzer]

Hi @grst, thanks for the reminder! Honestly it dropped off my radar since I have been busy with the LLM extension framework BioChatter, but I am still committed to doing this.

We are in fact planning several developments for making available a public (generic) API, and we have some projects that also created specific APIs on their own. I would be happy to have a meeting about specific requirements in the scirpy case.

[slobentanzer]

Hi @grst and others, just as a brief update, we will pursue this in the context of an ongoing immune receptor project. I can't state a concrete timeline, but we will address the problem. I will get in touch when it becomes acute. :)