Help needed in preprocessing scATAC-seq

[alirezaomidi]

Hi,

I want to preprocess some scATAC-seq datasets properly. I have seen episcanpy tutorials. I do binarization, variant peak selection and use regress_out method but still get poor results. Would you suggest me what to do in order to get better results? I have uploaded two of my preprocessing notebooks here and here.

@DaneseAnna

[DaneseAnna]

Hi,
Thank you for your feed back. You should avoid regress_out for now. Try increasing your number of features to 40k and use epi.pp.normalize_per_cell to correct for library size.
I hope it will help you.