Open christopher-vollmers opened 2 years ago
Hey Chris, happy to hear from you!
Northstar is a temperamental but very educational sandbox rather than a pipeline. Do you have time for a zoom one of these days so we can discuss synchronously?
If so I'll write you an email and we can arrange?
Hi Fabio,
I hope you and doing well (and still monitoring this repository). I'm trying to generate Atlases from the Tabula Sapiens h5ad files that you can download from figshare.
I want to use them to classify tissue-matched samples. I've gotten northstar to run through (the "subsample" version) but maybe classify 10% of new cells.
I was wondering what type of processing I should do before reading these files with annData and calling the northstar.average_atlas()/northstar.subsample_atlas() and northstar.Averages()/northstar.Subsample functions.
Also, what kind of preprocessing should I do on in let's say Seurat on a standard 10x run?
Would really appreciate the help! All other cell assignment pipelines I have tried are basically unusable.
Thanks, Chris