data-raw/gse167998-reference.r

In perishky/meffil: Efficient algorithms for DNA methylation

#' Defines cell type reference "blood gse167998"
#' 
#' Salas LA et al. (2022) Enhanced cell deconvolution of peripheral blood using DNA methylation for high-resolution immune profiling 

retrieve.gse167998 <- function(path) {
    source("geo.r")

    url <- "https://ftp.ncbi.nlm.nih.gov/geo/series/GSE167nnn/GSE167998/suppl/GSE167998_RAW.tar"
    filename <-  file.path(path, basename(url))
    download.file(url, filename)
    cat(date(), "Extracting files from GEO archive.\n")
    system(paste("cd", path, ";", "tar xvf", basename(filename)))
    unlink(filename)
    cat(date(), "Unzipping IDAT files.\n")
    system(paste("cd", path, ";", "gunzip *.idat.gz"))

    samplesheet <- meffil.create.samplesheet(path)	
    samplesheet <- samplesheet[, c("Sample_Name", "Slide", "Basename")]
    samples <- geo.samples("GSE167998")
    names(samples)[names(samples) == "geo_accession"] <- "Sample_Name"
    names(samples)[names(samples) == "description"] <- "CellType"
    samples <- merge(samples, samplesheet)

    samples
}

create.gse167998.reference <- function() {
    if (!requireNamespace("BiocManager", quietly = TRUE))
        install.packages("BiocManager")
    if (!requireNamespace("minfi", quietly = TRUE))
        BiocManager::install("minfi")
    if (!requireNamespace("IlluminaHumanMethylationEPICmanifest", quietly = TRUE))
        BiocManager::install("IlluminaHumanMethylationEPICmanifest")
    if (!requireNamespace("IlluminaHumanMethylationEPICanno.ilm10b4.hg19", quietly = TRUE))
        BiocManager::install("IlluminaHumanMethylationEPICanno.ilm10b4.hg19")

    require("minfi")
    require("IlluminaHumanMethylationEPICanno.ilm10b4.hg19")
    source("minfiEPIC.r")

    dir.create(path <- "GSE167998")
    on.exit(unlink(path, recursive=TRUE))
    samples <- retrieve.gse167998(path)
    
    RGset <- read.metharray.exp(targets = samples, force=TRUE)
    
    reference <- preprocessNoob(RGset)
    reference <- mapToGenome(reference)
    
    extracted.data <- extractFromRGSetEPIC(RGset)
    samplesheet <- colData(RGset)
    samplesheet <- merge(samplesheet, samples)
    sex <- getSex(reference, cutoff = -3)$predictedSex
    sex <- sign(sex=="F")+1
    reference <- normalizeFunnormEPIC(
        object=reference, extractedData=extracted.data,
        sex=sex, nPCs=10, verbose=F)
    M <- getMeth(reference)
    U <- getUnmeth(reference)
    
    ## Add cell type reference to meffil
    celltypes <- c("Neu", "Eos", "Bas", "Mono", 
               	   "Bnv", "Bmem", "CD4nv", "CD4mem", 
               	   "Treg", "CD8nv", "CD8mem", "NK", "MIX")
    cell.types <- celltypes[celltypes != "MIX"]
    stopifnot(all(cell.types %in% samplesheet$CellType))
    selected <- samplesheet$CellType %in% cell.types
    meffil.add.cell.type.reference(
        "blood gse167998",
        M[,selected], U[,selected],
        cell.types=samplesheet$CellType[selected],
        chip="epic",
        featureset="epic",
        ## specific.sites=IDOLOptimizedCpGs, # use of specific sites would require a license
        description="Adult blood reference of Salas et al. Nat Comms 2022",
        verbose=T)
}