In LTLA/CellTypeReferences: 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/dice/"
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", "dice.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="Human bulk RNA-seq data from DICE",
    description=paste(c(
        "Normalized expression values of 1561 bulk RNA-seq samples of sorted human immune cell populations from the Database of Immune Cell Expression (DICE).",
        "TPM normalized values for each cell type were downloaded from https://dice-database.org/downloads.",
        "Genes with no reads across samples were removed, and values were log~2~-normalized after a pseudocount of 1 was added.",
        "Samples were annotated to 5 main cell types (`label.main`):",
        "- B cells",
        "- Monocytes",
        "- NK cells",
        "- T cells, CD8+",
        "- T cells, CD4+",
        "",
        "These labels were further resolved to 15 fine cell types (`label.fine`):",
        "- B cells, naive",
        "- Monocytes, CD14+",
        "- Monocytes, CD16+",
        "- NK cells",
        "- T cells, memory TREG",
        "- T cells, CD4+, naive",
        "- T cells, CD4+, naive, stimulated",
        "- T cells, CD4+, naive Treg",
        "- T cells, CD4+, Th1",
        "- T cells, CD4+, Th1_17",
        "- T cells, CD4+, Th2",
        "- T cells, CD8+, naïve",
        "- T cells, CD8+, naïve, stimulated",
        "- T cells, CD4+, TFH",
        "- T cells, CD4+, Th17",
        "",
        "The fine types have also been mapped to the Cell Ontology (`label.ont`)."
    ), collapse="\n"),
    taxonomy_id="9606",
    genome=character(0), # dunno, it doesn't say.
    sources=list(
        list(provider="PubMed", id="30449622"),
        list(provider="ExperimentHub", id=norm.hub$ah_id),
        list(provider="ExperimentHub", id=coldata.hub$ah_id),
        list(provider="URL", id="https://dice-database.org/downloads", version="2019-09-04")
    ),
    maintainer_name="Jared Andrews",
    maintainer_email="[e-mail]"
)

Saving it to disk.

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

LTLA/CellTypeReferences documentation built on June 1, 2024, 12:12 p.m.