Multiplets after scrublet?

[rhlgyb]

Dear managers of scrublet,

Thank you for your time. I will try to be concise.

The gist is, I am curious, with my limited knowledge, what other intepretations or hypotheses I could make other than these cells being multiplets...I'm sure I used scrublet correctly.

I have used

• cellbender : to rid low quality cells filled with ambient RNA.
• (afterwards) : cellbender filtered h5 file -> Seurat count matrix -> subsetted seurat object with scrublet barcodes
• numerous combinations of normalization and batch correction methods
• tried removing some blacklist genes before processing the data

However, I still end up with a number of cells from a specific sample that expresses the marker genes of fibroblasts, immune cells, epithelial and endothelial cells; pretty much all of the major cell types.
(cells were used for this analysis : percent.mt < 10~20, etc, 500 < nFeatures < 6000, 1000 < UMI)

Thank you so much for your help,

R.

• (above) colors represent respective samples