is it possible to perform an integrated analysis of separated images?

[bugie19]

hallo everyone,

I have 3 separately different images with their individual CSV files obtained from CODEX staining method. The samples are timepoint of tumor stage (day 0, day10 and day20). Inspired from the integrated analysis of several single cell RNA sequencing dataset which is could be done in seurat package, Is it also possible to perform such an integrated analysis of those images in Scimap package?

your help i really appreciate in advance.

best regards,

Bugie

[ajitjohnson]

Hi @bugie19 Yes it is absolutely possible to do so.

If you are using sm.pp.mcmicro_to_scimap function to create the adata object, you can simply import all the images together like so:

all_images = ["\path\to\image_1.csv", "\path\to\image_2.csv", "\path\to\image_3.csv"]
adata = sm.pp.mcmicro_to_scimap (feature_table_path=all_images)

If you are manually creating the adata object, make sure you have a column called imageid in adata.obs that maps the cells to a unique ID for each image. The program automatically takes that into consideration for all downstream analyses.

[sailseem]

I have a follow-up query. Let's say I have 100 images, each with 20-30 markers, and I've constructed the scimap adata. However, when I execute UMAP, it doesn't delineate any distinct clusters. Instead, all the cells are grouped together. Do you have any advice on addressing this issue?