Hello,

I am analysing Affymetryx and Illumina data from the MACQ project for the moment. I used the following commands to read and preprocess (normalisation, background correction,..) the Affymetry CEL files:

files <- list.files('DONNEES-AFFYMETRIX/')
files <- paste0('DONNEES-AFFYMETRIX/',files)
rawfiles <- ReadAffy(filenames=files)
expressionset <- gcrma(rawfiles)

I don't know what commands to use to read AND preprocess ILLUMINA file (eg. ILM1A1.txt ) downloaded from NCBI:

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE5350 :

I started with the same commands as for the Affymetrix cell files;namely

files <- list.files('DONNEES-ILLUMINA/')
files <- paste0('DONNEES-ILLUMINA/',files)

and tried to read the files with read.ilmn but with no success; the file content looks like this:

getID   ILM_1_A1    BEAD_STDEV-A1   Avg_NBEADS-A1   Detection-A1
GI_10047089-S   59.4    3.3 45  0.8431114
GI_10047091-S   90.4    3.1 50  0.99802241
GI_10047093-S   539.8   13.5    43  1
GI_10047099-S   563.6   21.2    32  1

I read about the existence of "rsn" "ssn" "lumiR" and "lumiR.batch" for preprocessing( normalisation...) but couldn't use them as I can't read the data to start with. I have spent quite a long time searching the internet for answers but couldn't find any. Could you perhaps help me with that?? It would be really great!

The MAQC Illumina microarray data deposited on GEO do not have the original format. So the default parameter values of read.ilmn function won't work. Below is an example command to read in one of the files:

library(limma)
x <- read.ilmn("ILM_1_A1.txt",probeid="TargetID",expr="ILM_")

There isn't a ready-made solution for reading lots of single-array Illumina data files, as from GEO, because Illumina raw data files normally have data from all the microarrays in one file. The beauty of R however is that you can easily solve problems like this in a few lines of programming. In this case, just read all the files separately with Wei's code and store them in a list:

library(limma)
Input <- list()
for (f in files) Input[[f]] <- read.ilmn(f, probeid="TargetID", expr="ILM_")

Then cbind them together into one EListRaw object:

x <- do.call(cbind, Input)

Then normalize:

y <- neqc(x)

Now you're ready to analyze. y is an EList object and the normalized log-expression values are in y$E. Most limma functions will operate on y as an object.

The above code runs in a few seconds on my PC.