search

BayraktarLab / cell2location

Comprehensive mapping of tissue cell architecture via integrated single cell and spatial transcriptomics (cell2location model)
https://cell2location.readthedocs.io/en/latest/
Apache License 2.0
321 stars 58 forks source link

Question: cell proportion, cell abundance and spatial clustering #342

Closed dyinboisry4u closed 10 months ago

dyinboisry4u commented 10 months ago

Hi, @vitkl, thanks to develop a great tool, it helps me a lot, I have several questions about c2l downstream analysis:

  1. In c2l tutorial, we use cell abundance estimated by c2l as downstream clustering input. And you suggest Normalising to the total per spot doesn't simplify comparisons of cell abundance between cell types in any way. (https://github.com/BayraktarLab/cell2location/issues/151#issuecomment-1140102241). But in some research(https://doi.org/10.1038/s41586-022-05060-x), I noticed the author use c2l cell proportion with performing isometric log ratios transformation (https://github.com/saezlab/visium_heart/issues/7#issuecomment-1257764447) as clustering input. I'm really confused if I can use cell proportion per spot for downstream analysis.
  2. Could I use some spatial clustering methods to detect spatial domains, and then assign c2l cell abundance for each domains to illustrate the cell composition of each domain(something like domain - cell abundance dotplot)?
dyinboisry4u commented 10 months ago

I got it in c2l paper