In drisso/scRNAseq: Collection of Public Single-Cell RNA-Seq Datasets

library(BiocStyle)
knitr::opts_chunk$set(error=FALSE, message=FALSE, warning=FALSE)

Downloading the count data

We obtain a single-cell RNA sequencing dataset of the mouse hypothalamus from @romanov2017molecular. Counts for endogenous genes are available from the Gene Expression Omnibus using the accession number GSE74672. We download and cache it using the r Biocpkg("BiocFileCache") package.

library(BiocFileCache)
bfc <- BiocFileCache("raw_data", ask = FALSE)
base.url <- file.path("ftp://ftp.ncbi.nlm.nih.gov/geo/series",
    "GSE74nnn/GSE74672/suppl")
count.file <- bfcrpath(bfc, file.path(base.url,
    "GSE74672_expressed_mols_with_classes.xlsx.gz"))

We load them into memory.

library(R.utils)
tmp.file <- tempfile(fileext=".xlsx")
gunzip(count.file, destname=tmp.file, remove=FALSE)

# WARNING: this requires >20 GB of memory!
library(readxl)
all.data <- read_xlsx(tmp.file, col_names=FALSE)
dim(all.data)

Extracting the metadata

The first 9 rows correspond to metadata and need to be extracted.

coldata <- all.data[1:9,]
colfields <- coldata[,1,drop=TRUE]
coldata <- coldata[,-1]

coldata <- t(coldata)
coldata <- data.frame(coldata, stringsAsFactors=FALSE)
rownames(coldata) <- NULL

library(S4Vectors)
coldata <- DataFrame(coldata)
colnames(coldata) <- colfields

for (i in 4:9) {
    coldata[,i] <- as.numeric(coldata[,i])
}
coldata

The next three rows are empty, so we just skip them. We convert the remaining counts into a sparse matrix.

raw.counts <- all.data[-(1:12),]
counts <- as.matrix(raw.counts[,-1])

library(Matrix)
counts <- as(counts, "dgCMatrix")
rownames(counts) <- raw.counts[,1,drop=TRUE]
colnames(counts) <- NULL
dim(counts)

Saving to file

We now save all of the components to file for upload to r Biocpkg("ExperimentHub"). These will be used to construct a SingleCellExperiment on the client side when the dataset is requested.

path <- file.path("scRNAseq", "romanov-brain", "2.0.0")
dir.create(path, showWarnings=FALSE, recursive=TRUE)
saveRDS(counts, file=file.path(path, "counts.rds"))
saveRDS(coldata, file=file.path(path, "coldata.rds"))