log2TransformPositive: Log2 transform of numbers >1.

In blima: Tools for the preprocessing and analysis of the Illumina microarrays on the detector (bead) level

Description

Transformation function are popular in beadarray package. Here this is similar concept. This function allow user to perform log transformation before doing t-tests.

Usage

log2TransformPositive(x)

Arguments

x	Number to transform.

Value

This function returns logarithm of base 2 for numbers >=1 and zero for numbers <1.

Author(s)

Vojt<c4><9b>ch Kulvait

Examples

if(require("blimaTestingData") && require("illuminaHumanv4.db") && interactive())
{
    #To perform background correction, quantile normalization and then bead level t-test on log data run. Vst is not performed in this scheme. Top 10 probes is then printed according to certain measure.
    data(blimatesting)
    #Prepare logical vectors corresponding to conditions A(groups1Mod), E(groups2Mod) and both(c).
    groups1 = "A";
    groups2 = "E";
    sampleNames = list()
    groups1Mod = list()
    groups2Mod = list()
    c = list()
    for(i in 1:length(blimatesting))
    {
        p = pData(blimatesting[[i]]@experimentData$phenoData)
        groups1Mod[[i]] = p$Group %in% groups1;
        groups2Mod[[i]] = p$Group %in% groups2;
        c[[i]] = p$Group %in% c(groups1, groups2);
        sampleNames[[i]] = p$Name
    }
    #Background correction and quantile normalization followed by testing including log2TransformPositive transformation.
    blimatesting = bacgroundCorrect(blimatesting, normalizationMod =c, channelBackgroundFilter="bgf")
    blimatesting = nonPositiveCorrect(blimatesting, normalizationMod=c, channelCorrect="GrnF",  channelBackgroundFilter="bgf", channelAndVector="bgf")
    blimatesting = quantileNormalize(blimatesting, normalizationMod=c, channelNormalize="GrnF", channelOutput="qua", channelInclude="bgf")
    beadTest <- doTTests(blimatesting, groups1Mod, groups2Mod,
            transformation=log2TransformPositive, quality="qua", channelInclude="bgf")
    symbol2address <- merge(toTable(illuminaHumanv4ARRAYADDRESS), toTable(illuminaHumanv4SYMBOLREANNOTATED))
    symbol2address <- symbol2address[,c("SymbolReannotated", "ArrayAddress") ]
    colnames(symbol2address) <- c("Symbol", "ArrayAddressID")
    beadTest = merge(beadTest, symbol2address, by.x="ProbeID", by.y="ArrayAddressID")
    beadTestID = beadTest[,c("ProbeID", "Symbol")]
    beadTestFC = abs(beadTest[,"mean1"]-beadTest[,"mean2"])
    beadTestP = beadTest[,"adjustedp"]
    beadTestMeasure = (1-beadTestP)*beadTestFC
    beadTest = cbind(beadTestID, beadTestMeasure)
    colnames(beadTest) <- c("ArrayAddressID", "Symbol", "difexBL")
    sortBL = sort(-beadTest[,"difexBL"], index.return=TRUE)$ix
    beadTop10 = beadTest[sortBL[1:10],]
    print(beadTop10)
}else
{
    print("To run this example, please install blimaTestingData package from bioconductor by running BiocManager::install('blimaTestingData') and illuminaHumanv4.db by running BiocManager::install('illuminaHumanv4.db').");
}

blima documentation built on Nov. 8, 2020, 8:15 p.m.