Not able to filter out the empty spot using the spatial spot level quality control

[ashishjain1988]

Hi,

I am trying to carry out the spot level quality control for the library size, unique genes, and mitochondrial percent using the localOutliers function on my 10X Visium spatial transcriptomics data. I found that it was able to filter out few of spots based but not able to filter out empty spots in the slide (left region). Is there a way that we can filter out empty spots using SpotSweeper? I also tried Hangnail artifact removal function but it was filtering out other spots not the empty ones. Can you please help?

Regards,
Ashish Jain

[stephaniehicks]

hi @ashishjain1988 thanks for your interest! That looks like a dry edge artifact. We are actively working on adding this functionality right now (#22 and #23 (comment)). We hope to have it implemented within the next few weeks. I'll leave the issue for now and will loop back as soon as it's available. Hope this helps!

[MicTott]

Hi @ashishjain1988,

When you say these are "empty" do you mean that they have lower library size than the rest of the tissue? SpotSweeper's localOutlier function works by detecting spatial outliers. Since the left side of your data is are all of lower library size, none of them are local outliers, and it's correctly ignoring them.

Are you sure that this area is of low quality? Many biological domains naturally show lower library size. If it is indeed low quality / non-biological, we are working on an edge dry spot detection function that should be able to accurately detect it. In the mean time, you may try using fixed thresholding to remove those spots, similar to snRNA-seq based methods. https://bioconductor.org/books/3.14/OSCA.basic/quality-control.html

[ashishjain1988]

Hey @MicTott , yes these seems to be dry/empty spots as there is no expression of any gene in those spots. Those spots are filtered out if we used fixed parameters for quality controls.