search

ncborcherding / utility

Collection of Tumor-Infiltrating Lymphocyte Single-Cell Experiments with TCR
Other
81 stars 7 forks source link

any plan for h5ad (scanpy) version #1

Closed M0hammadL closed 2 years ago

M0hammadL commented 2 years ago

Hi

thanks for sharing the data and letting the world use it for science!
Would it be possible to also release the h5ad version of the file readable via scanpy? many people use scanpy and would be great if you release it.

Best Mo

ncborcherding commented 2 years ago

Hey Mo,

Absolutely, that is my goal for the next data update. Do you prefer h5 or loom?

Thanks, Nick

M0hammadL commented 2 years ago

h5ad is the default for scanpy:

https://anndata.readthedocs.io/en/latest/anndata.AnnData.html

ncborcherding commented 2 years ago

Ok great - thanks for the info. I will post an update here when I get it converted.

Nick

grst commented 2 years ago

Just read on twitter you are planning a scanpy-ecosystem version.

You may want to consider loading the TCR data with scirpy: https://icbi-lab.github.io/scirpy/latest/tutorials/tutorial_io.html

Essentially, scirpy stores TCR data in adata.obs using AIRR-compliant column names. LMK if you have any questions or need other input.

ncborcherding commented 2 years ago

Thanks for the suggestion - will definitely look into it. Please feel free to make others as I am not as familiar with python as R!

grst commented 2 years ago

For full disclosure I need to add that I'm the main developer of scirpy.

Another very cool tool in the Python world is conga which focuses on the integration of TCR with gene expression data. We are working on adding a conga wrapper to scirpy.

There's also dandelion but it is mostly specialized on BCR data.

M0hammadL commented 2 years ago

thanks, @grst @ncborcherding it would be really beneficial for further usage of the data and opening a new community!

PS: @grst is great, I use it all the time, keep on a good work!

ncborcherding commented 2 years ago

Thanks I appreciate the suggestion and I will have to check out scirpy more - I might ping you if I have questions - full disclosure, I wrote an R package for single-cell immune repertoire profiling - scRepertoire and am deeply interested in how people are using the interaction of TCR/GEX - which CONGA is definitely a great example.

thanks @M0hammadL and @grst for the suggestions!!

ncborcherding commented 2 years ago

Finally got .h5ad up for the sequencing runs, so I am going to close this - I have yet to merge/integrate all cells into a single object.

@grst - still looking at making a formal python pipeline using scanpy/scirpy. I think that will be a future goal!