GSE62944_TCGA_Rusbread_featureCounts.R

## Imports: GEOquery, Biobase

acc <- "GSE62944"
if (!file.exists(acc))
    GEOquery::getGEOSuppFiles(acc)
setwd(acc)

clinvar <- local({
    message("clinvar")
    fl <- "GSE62944_TCGA_20_420_Clinical_Variables_7706_Samples.txt.gz"
    data <- scan(fl, what=character(), sep="\t", quote="")
    m <- matrix(data, 7707)
    dimnames(m) <- list(m[,1], m[1,])
    df <- as.data.frame(m[-1, -1])

    fl <- "GSE62944_TCGA_20_CancerType_Samples.txt.gz"
    ct <- read.delim(fl, header=FALSE,
                     colClasses=c("character", "factor"),
                     col.names=c("sample", "type"))
    idx <- match(rownames(df), ct$sample)
    stopifnot(!anyNA(idx))
    df$CancerType <- ct$type[idx]
    df
})


counts <- local({
    message("counts")
    fl <- "GSM1536837_TCGA_20.Illumina.tumor_Rsubread_FeatureCounts.txt.gz"
    if (!file.exists(fl))
        untar("GSE62944_RAW.tar", fl)
    m <- scan(fl, what=character(), sep="\t", quote="")
    m <- matrix(m, 7707)
    dimnames(m) <- list(m[,1], m[1,])
    m <- t(m[-1, -1])
    mode(m) <- "integer"
    m
})

adf <- Biobase::AnnotatedDataFrame(clinvar)
eset <- Biobase::ExpressionSet(counts, adf)