Micah's RNA+ADT+HTO update using UM5 and Michael porjects as templates

[micahpf]

Note that I wasn't able to finish all the planned changes before I left Emory.

I left off partway through the integration and annotation runfile. You can use the UM5 subsetting runfiles/formatfiles and michael integration+annotation runfiles/formatfiles as references to finish that.

The next step would be to add the pseudobulk script from the Michael project. Fingers crossed this slots in seamlessly with little changes required.

Also note that there is a bug in the processing runfile at seurat object creation due to package conflicts. I think this could be resolved by updating R to the latest version with a bioconductor base docker image.

Good luck!

[akcorut]

This fails with Error in .subscript.2ary(x, , j, drop = TRUE) : subscript out of bounds

[akcorut]

scRNAseq_template/analysis_scripts/01-processing-template_ADT+soupX.runfile.Rmd

Lines 513 to 517 in 8484950

	# Add capID to front of cell barcode to prevent collisions when merging across clusters
	sc <- lapply(sc, function(x) {
	x$adjusted_counts@Dimnames[[2]] <- paste0(capID, "_", x$adjusted_counts@Dimnames[[2]])
	return(x)
	})

This error seems to be caused by a bug in the above part where all capIDs are replaced by the last capID.

[akcorut]

Below is a fix for this issue.

Suggested change

	s.split.unfilt <- lapply(names(s.split.unfilt), function(capID) { #s.split.unfilt
	s.split.unfilt[[capID]]@assays$RNA@layers$counts <-
	sc[[capID]]$adjusted_counts[,colnames(s.split.unfilt[[capID]])]
	return(s.split.unfilt[[capID]])
	})
	# Add capID to front of cell barcode to prevent collisions when merging across clusters
	for (capID in names(sc)) {
	sc[[capID]]$adjusted_counts@Dimnames[[2]] <- paste0(capID, "_", sc[[capID]]$adjusted_counts@Dimnames[[2]])
	}
	## Initialize a flag to track if all columns are present
	all_present <- TRUE
	## Check for missing columns in each capID
	for (capID in names(sc)) {
	missing_cols <- setdiff(colnames(s.split.unfilt[[capID]]), colnames(sc[[capID]]$adjusted_counts))
	if (length(missing_cols) > 0) {
	warning(paste0("The following columns are missing in adjusted_counts for ", capID, ": ", paste(missing_cols, collapse = ", ")))
	all_present <- FALSE # Set flag to FALSE if any missing columns are found
	} else {
	message(paste0("All columns present for ", capID))
	}
	}
	# Replace cellranger filtered counts with soupx adjusted counts of filtered cells
	if (all_present) {
	## Execute lapply only if all columns are present in all capIDs
	message("All columns are present. Replacing cellranger filtered counts with soupx adjusted counts.")
	s.split.unfilt <- lapply(names(s.split.unfilt), function(capID) {
	s.split.unfilt[[capID]]@assays$RNA@layers$counts <-
	sc[[capID]]$adjusted_counts[, colnames(s.split.unfilt[[capID]])]
	return(s.split.unfilt[[capID]])
	})
	names(s.split.unfilt) <- names(sc)
	} else {
	# if all columns are not present, stop the script
	stop("Not all columns are present. The replacement was not executed.")
	}

[akcorut]

There are bugs in write_rds. paste0() is missing when getDate() is used.

Suggested change

	write_rds(s.split.unfilt, here(paste0("saved_rds/report-", report_number), "s.split.unfilt-", getDate(), ".rds")) # for filtering plots
	write_rds(s.demux, here(paste0("saved_rds/report-", report_number), "s.demux-", getDate(), ".rds"))
	write_rds(s.filtADT.split, here(paste0("saved_rds/report-", report_number), "s.filtADT.split-", getDate(), ".rds"))
	write_rds(s.filtADT, here(paste0("saved_rds/report-", report_number), "s.filtADT-", getDate(), ".rds")) # For downstream processing
	```
	write_rds(singler_out, here(paste0("saved_rds/report-", report_number), paste0("singler_out-", getDate(), ".rds")))
	write_rds(s.split.unfilt, here(paste0("saved_rds/report-", report_number), paste0("s.split.unfilt-", getDate(), ".rds"))) # for filtering plots
	write_rds(s.demux, here(paste0("saved_rds/report-", report_number), paste0("s.demux-", getDate(), ".rds")))
	write_rds(s.filtADT.split, here(paste0("saved_rds/report-", report_number), paste0("s.filtADT.split-", getDate(), ".rds")))
	write_rds(s.filtADT, here(paste0("saved_rds/report-", report_number), paste0("s.filtADT-", getDate(), ".rds"))) # For downstream processing