Error: SPOTlight command in R

Question

[Hien]

Hi everyone,

Do anyone help me to figure out my error in SPOTlight command? Tutorials and outsourcing data for SPOTlight are not enough for me to fix my errors.

I have two Seurat objects: one is a Seurat object from https://www.gutcellatlas.org/ for reference with annotating cell lines, and the other is a Seurat object from my sample (Visium spatial gene expression).

spotlight_ls <- SPOTlight(
  x = seurat_cap3,
  y = data_sc,
  groups = "category", # Variable in sc_seu containing the cell-type annotation
  mgs = test, # Dataframe with the marker genes
  n_top = NULL,
  gene_id = "gene",
  weight_id = "weight",
  group_id = "type",
  gene_id = "gene"
)

It always shows the error:

Error in SPOTlight(x = seurat_sc, y = data_sc, groups = "category", mgs = test,  : formal argument "gene_id" matched by multiple actual arguments

sessionInfo(package = NULL)
R version 4.3.1 (2023-06-16 ucrt)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 19045)

Thank you so much!

Answer 1

[Matthias Zepper]

That is an easy fix, but of course easy to overlook - you accidentally provided the argument gene_id twice.

[Hien]

Hi, Thank you for your indication! I have fixed the command as removing gene_id = gene and the title of mgs(cluster_markers_all) following "gene", "type", and "weight".

nmf_mod_ls = SPOTlight( x = seurat_sc, 
                       y = data_sc@assays$Spatial@counts, 
                       group = "category", 
                       mgs = cluster_test,
                       weight_id = "weight",
                       group_id = "type",
                       gene_id = "gene")

However, again, it shows the different error

Error in trainNMF(x = x, y = y, groups = groups, mgs = mgs, n_top = n_top,  : 
  is.null(groups) | length(groups) == ncol(x) is not TRUE

Right now, I have not indicated how to fix it :(

[Matthias Zepper]

Unfortunately, I have no experience with SPOTlight. The previous error was a generic R issue, so I could help nonetheless, but this one is clearly package specific.

However, is.null(groups) | length(groups) == ncol(x) is not TRUE tests that either no groups are provided or the length of the group vector corresponds to the number of columns in x. So try to run ncol(seurat_sc) and provide a vector of exactly this length to group. Assuming your seurat_sc would only comprise 4 columns/cells, something like group = c("A","B","A","C") could work?