How to choose the library normalization size in single cell RNA-seq data Preprocessing?

Question

[Dan]

Hello,

I used scanpy.pp.normalize_total function in SCANPY package to normalize the library to the same size. (https://scanpy.readthedocs.io/en/stable/generated/scanpy.pp.normalize_total.html#scanpy.pp.normalize_total) I found the target_sum is different among kinds of literatures, some used 10K, some used CPM, and very rarely used median library size. I am not sure how to choose the library normalization size?

Thanks a lot