Generate a fraction of immune cell between two condition from single cell RNAseq

[synatkeamsk]

Dear respected professor,

Do you know whether I can use Seurat generate fractions of immune cells from total cells as in figure (e) of attached image between two condition. Could not find any script to do that. Really appreciate your help.

Regards,

[Zig-E-Stardust]

You can try the following code (make sure to have Seurat loaded, as well as to have the following packages "foreach" and "openxlsx" installed; in addition, change the names where needed - e.g., if your main object is not called "combined", replace "combined" with whatever name you gave that object) - tell me if it worked for you:

library(foreach)
combined$annotation_mouse <- paste(combined$annotation, combined$mouse, sep = "_")
combined$annotation_group <- paste(combined$annotation, combined$group, sep = "_")

Idents(combined) <- "annotation"
levels(combined)

# create a vector of cluster names
clusters <- levels(combined)

# create an empty vector to store the results
PCT <- vector()

# loop over the clusters and calculate the percentage for each mouse
PCT <- foreach(c = clusters, .combine = c) %do% {
  # create a vector of mouse names
  mice <- c("WT1", "WT2", "WT3", "WT4", "KO1", "KO2", "KO3", "KO4")
  
  # create an empty vector to store the percentages for each cluster
  PCT_c <- vector()
  
  # loop over the mice and calculate the percentage for each cluster
  for (m in mice) {
    # create a variable for the subset_mouse column name
    subset_mouse <- paste0(c, "_", m)
    
    # calculate the percentage:
    PCT_m <- (sum(combined$annotation_mouse == subset_mouse) / sum(combined$mouse == m)) * 100
    
    # assign a name to the percentage
    names(PCT_m) <- paste0(m, "_", c)
    
    # append the percentage to the vector
    PCT_c <- c(PCT_c, PCT_m)
  }
  
  # return the vector of percentages for each cluster
  return(PCT_c)
}

# save as an Excel (.xlsx) file
PCT <- as.data.frame(PCT)
library(openxlsx)
write.xlsx(PCT, "cluster percentages.xlsx", rowNames = TRUE)


# Check that the results are correct, but comparing the results of one cluster to those found in the Excel file:
Cluster_XYZ_PCT <- c(((sum(combined$annotation_mouse=="Cluster_XYZ_WT1")/sum(combined$mouse=="WT1"))*100),
                  ((sum(combined$annotation_mouse=="Cluster_XYZ_WT2")/sum(combined$mouse=="WT2"))*100),
                  ((sum(combined$annotation_mouse=="Cluster_XYZ_WT3")/sum(combined$mouse=="WT3"))*100),
                  ((sum(combined$annotation_mouse=="Cluster_XYZ_WT4")/sum(combined$mouse=="WT4"))*100),
                  ((sum(combined$annotation_mouse=="Cluster_XYZ_KO1")/sum(combined$mouse=="KO1"))*100),
                  ((sum(combined$annotation_mouse=="Cluster_XYZ_KO2")/sum(combined$mouse=="KO2"))*100),
                  ((sum(combined$annotation_mouse=="Cluster_XYZ_KO3")/sum(combined$mouse=="KO3"))*100),
                  ((sum(combined$annotation_mouse=="Cluster_XYZ_KO4")/sum(combined$mouse=="KO4"))*100))

Mouse <- c("WT1", "WT2", "WT3", "WT4", "KO1", "KO2", "KO3", "KO4")

Cluster_XYZ_Percentages <- data.frame(Mouse, Cluster_XYZ_PCT)