In LTLA/celldex: Index of Reference Cell Type Datasets

From ExperimentHub to gypsum

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

Pulling assets from ExperimentHub.

library(ExperimentHub)
hub <- ExperimentHub()
prefix <- "celldex/mouse.rnaseq/"
norm.hub <- hub[hub$rdatapath==paste0(prefix, "1.0.0/logcounts.rds")]
stopifnot(nrow(norm.hub) == 1L)
norm <- norm.hub[[1]]
str(norm)
coldata.hub <- hub[hub$rdatapath==paste0(prefix, "1.0.0/coldata.rds")]
stopifnot(nrow(coldata.hub) == 1L)
coldata <- coldata.hub[[1]]
coldata

Attaching ontology mappings.

path <- system.file("mapping", "mouse_rnaseq.tsv", package="celldex", mustWork=TRUE)
src <- read.delim(path, header=FALSE, stringsAsFactors=FALSE)

m <- match(coldata$label.fine, src[,1])
stopifnot(all(!is.na(m))) # sanity check

matched <- src[m, 2]
matched[matched==""] <- NA_character_
coldata$label.ont <- matched
coldata

Assembling some metadata.

meta <- list(
    title="Bulk RNA-seq data of sorted mouse cell populations",
    description=paste(c(
        "Log-normalized expression values of 358 bulk RNA-seq samples of sorted cell populations in GEO.",
        "This dataset was contributed by the Benayoun Lab, who identified, downloaded and processed data sets on GEO corresponding to sorted cell types (Benayoun et al., 2019).",
        "Samples were annotated to 18 main cell types (`label.main`):",
        "- Adipocytes",
        "- Astrocytes",
        "- B cells ",
        "- Cardiomyocytes",
        "- Dendritic cells",
        "- Endothelial cells",
        "- Epithelial cells",
        "- Erythrocytes",
        "- Fibroblasts",
        "- Granulocytes",
        "- Hepatocytes",
        "- Macrophages",
        "- Microglia",
        "- Monocytes",
        "- Neurons",
        "- NK cells",
        "- Oligodendrocytes",
        "- T cells",
        "",
        "These were resolved further into 28 subtypes (`label.fine`), which have also been mapped to the Cell Ontology (`label.ont`)."
    ), collapse="\n"),
    taxonomy_id="10090",
    genome="MGSCv37",
    sources=list(
        list(provider="PubMed", id="30858345"),
        list(provider="URL", id="https://github.com/BenayounLaboratory/Mouse_Aging_Epigenomics_2018/tree/master/FigureS7_CIBERSORT/RNAseq_datasets_for_Deconvolution/2017-01-18"),
        list(provider="PubMed", id="30643263"),
        list(provider="GitHub", id="dviraran/SingleR", version="adc4a0e4d5cfa79db18f3821f51a02cbd6484710"),
        list(provider="ExperimentHub", id=norm.hub$ah_id),
        list(provider="ExperimentHub", id=coldata.hub$ah_id)
    ),
    maintainer_name="Friederike Dündar",
    maintainer_email="[e-mail]"
)

Saving it to disk.

library(celldex)
path <- "2024-02-26_output"
saveReference(norm, coldata, path, meta)

LTLA/celldex documentation built on June 3, 2024, 4:53 p.m.