scverse / scanpy

Single-cell analysis in Python. Scales to >1M cells.
https://scanpy.readthedocs.io
BSD 3-Clause "New" or "Revised" License
1.92k stars 602 forks source link

pbmc68k_reduced raw counts #3224

Open MartaBenegas opened 2 months ago

MartaBenegas commented 2 months ago

Hi!

I was playing with the pbmc68k_reduced dataset and I saw that it has decimal values even in the "raw" matrix:

>>> adata.raw.X.data
array([2.177, 2.177, 2.544, ..., 1.142, 2.255, 1.142], dtype=float32)

which transformation has been applied? shouldn't the "raw" group contain unprocessed counts? is there a standard for this? Ideally, I would like to know where to find the raw, unprocessed, counts everytime I see an anndata or h5ad file.