R: Gene Matrix (.Txt File) To Boxplot And Rle, Nuse Plots

Question

[michaelsbreen1]

Hi all,

I am analyzing microarray data. I am using a new normalization approach which outputs into a gene matrix text file.

Normally I take raw CEL files and use "ReadAffy" function, plot a raw boxplot, normalize with RMA, then plot a normalized boxplot following I write it to a gene matrix, something like this:

raw <- ReadAffy(filenames = sampleNames(pd), phenoData = pd, sampleNames = pd$SampleName)
boxplot(raw)
normalized <- gcrma(raw)
boxplot(normalized)
matrix <- exprs(normalized)
write.table(matrix, file = "geneMatrix.txt")
PsetNormalized <- fitPLM(raw, background.method=”GCRMA”)
# Plots the RLE graph
Mbox(PsetNormalized)
# Plots the NUSE graph
boxplot(PsetNormalized)

Currently, my normalization approach takes CEL files and outputs then directly into a log2 value gene matrix. Is there a function I can use to take a gene matrix with log2 values and create such box plots and later RLE and NUSE plots?

Any suggestions are grateful.

Thank you,

Michael

Answer 1

[Sean Davis]

If you have a matrix of values, you can plot a boxplot easily:

# an example:
m = matrix(rnorm(100000),nc=10)
boxplot(m)

If you have probeset-level data rather than gene-level data, I think you should be able to reproduce an RLE plot. As for NUSE, that will not be possible without probe-level data. That said, is there a reason not to use the .CEL files to produce the NUSE and RLE plots? These plots should still useful for QC of the arrays regardless of the normalization method.