Hi,

I'm analyzing a snRNA-seq dataset consisting of 43 integrated human brain samples across 2 brain regions. During the subtype annotation I noticed some clear signs of the presence of ambient RNA.

The DecontX ambient RNA removal tool seems to be the best fit for our data as I would like to keep my cell type annotations. Unfortunately I'm a bit stuck. Joining the counts of the samples does not seem to work (the matrix gets too big) and the memory usage is enormous. How would you propose to tackle this problem? Do I need to split the samples again, loop the code over every single one and then re-integrate everything?

Thanks in advance!