TM_Process_smartseq2.R

# R3.4
require("tibble")
require("scater")
require("SingleCellExperiment")

files <- Sys.glob("FACS/*-counts.csv.gz")
ann <- read.table("FACS_annotations.csv.gz", sep=",", header=TRUE)
meta <- read.table("FACS_metadata.csv.gz", sep=",", header=TRUE)
rownames(meta) <- meta[,1]

for( f in files ) {
	dat = read.delim(f, sep=",", header=TRUE)
	rownames(dat) <- dat[,1]
	dat <- dat[,-1]
	cellIDs <- colnames(dat)
	cell_info <- strsplit(cellIDs, "\\.")
	Well <- lapply(cell_info, function(x){x[1]})
	Well <- unlist(Well)
	Plate <- unlist(lapply(cell_info, function(x){x[2]}))
	Mouse <- unlist(lapply(cell_info, function(x){x[3]}))
	this_ann <- ann[match(cellIDs, ann[,1]),]
	celltype <- this_ann[,3]
	this_meta <- meta[Plate,]
	name <- this_meta[1,3]
	cell_anns <- data.frame(mouse = Mouse, well=Well, cell_type1=celltype, tissue=this_meta[,3], subtissue=this_meta[,4], FACS=this_meta[,5], sex=this_meta[,6])
	rownames(cell_anns) <- colnames(dat)
	#cell_anns <- new("AnnotatedDataFrame", data=cell_anns)
	gene_anns<- data.frame(feature_symbol=rownames(dat));
	#sceset <- scater::newSCESet(countData=as.matrix(dat), phenoData=cell_anns)
	sceset <- SingleCellExperiment(assays = list(counts = as.matrix(dat)), colData=cell_anns, rowData=gene_anns)
	saveRDS(sceset, file=paste(as.character(name), "FACS.rds", sep="_"))
}