Filtering?

[idalarsson]

Hi! I tried running drug2cell on my single cell data and it works fine, but in the resulting adata-object only around 3000 cells are kept, even though around 30 000 cells go into the analysis from the original adata object. Is there some filtering going on in the d2c.score() function that can explain this? And is there a parameter to tune this so more cells are kept? Thanks!

[ktpolanski]

There is nothing in d2c.score() that does any sort of cell filtering. I tried for a fresh install and ran the following minimal example:

import scanpy as sc
import drug2cell as d2c

adata = sc.datasets.pbmc3k_processed()
d2c.score(adata, use_raw=True, targets = {"testing":["LINC00115"]})

Both adata and adata.uns["drug2cell"] have 2638 cells. 2620 of adata.uns["drug2cell"]'s cells have zero counts but they are there!