for an ongoing collaborative project, i want to analyze a transcriptomic dataset, which is not deposited in GEO, but in ArrayExpress:
So, through the R package arrayexpress, i tried initially to download the relative raw data:
rawset = ArrayExpress("E-MTAB-6326")
NChannelSet (storageMode: lockedEnvironment)
assayData: 62976 features, 119 samples
element names: E, Eb
US84600244_253949426813_S01_GE1_107_Sep09_1_4.txt (119 total)
varLabels: Source.Name Characteristics.organism. ...
Factor.Value.RIDD.score. (29 total)
varMetadata: labelDescription channel
featureNames: 1 2 ... 62976 (62976 total)
fvarLabels: Row Col ... SystematicName (5 total)
experimentData: use 'experimentData(object)'
However, my main concerns are the following:
1) Based on the manual of ArrayExpress, if the returned class object is an NChannelSet, then the experiment is a two-colour experiment-
but, when i checked the relative "Sample and data relationship" txt file, the column Label only reports the Cy3 label-
thus, this platform is indeed a two colour, or i have understood something wrong here ? Also this is evident, from the phenodata info from the downloaded object:
1428:Cy3 2948:Cy3 3174:Cy3 3178:Cy3 3183:Cy3 3280:Cy3
1 1 1 1 1 1
Just for simplicity, the platform is
the Agilent SurePrint G3 Human GE v2 8x60k Microarray (Array Design A-MEXP-2320)
2) In conjuction with my previous question, how should i continue with downstream analysis ? My main goals is primarily test a small gene signature on how the data are clustered, as also to perform a DE-analysis; however, how i should proceed with the NChannelSet ? in order to utilize limma for normalization , filtering etc ?
Thank you in advance,