negative cluster label

[FrenchSauce]

Hi!

I noticed that there is a '-1' cluster label generated after clustering (both with Leiden and Paris) it doesn't contain much cells (4 out of 50k in the test I run), I was just wondering whether it is a bug or something expected, I haven't found any mention of this in the documentation

clusters = ds_sample.cells.fetch_all('RNA_leiden_cluster') np.unique(clusters)

output:
array([-1, 1, 2, 3, 4, 5])

Thanks!

[parashardhapola]

Hi @FrenchSauce

fetch_all returns the data for all the cells, even those that were filtered out. Since the filtered out cells were not used for clustering they were assigned a value of -1. To ignore the data of filtered out cells you can use fetch instead of fetch_all.

Hope this helps.