Rather than starting at RMA, I'd like to load my Affy data into an expression set in R and use limma to do the full, proper fold change analysis. (My understanding is that RMA is quantile normalized, so fold change analysis won't be accurate.) My goal is to go from CEL files to the limma top table. Bioconductor has posted a very helpful page describing this (http://www.bioconductor.org/help/workflows/arrays/).
However, I don't have a clue what the right way is to build the phenoData object. I have seen very few examples of what the object even looks like. Here is the most detailed I have seen, located here: http://127.0.0.1:20212/library/Biobase/doc/ExpressionSetIntroduction.pdf.
gender type score A Female Control 0.75 B Male Case 0.40 C Male Control 0.73 D Male Case 0.42 E Female Case 0.93 F Male Control 0.22
But the way you are supposed to make this is using other components I don't understand. Code below is from the aforementioned Bioconductor page.
## import "phenotype" data, describing the experimental design phenoData <- read.AnnotatedDataFrame(system.file("extdata", "pdata.txt", package="arrays"))
Would it be okay if I just read a tab-delimited table into R? Like this hypothetical text object:
treatment A sunshine B deadly poison C sunshine D deadly poison E sunshine F deadly poison
...Because I understand how to do that, but not the phenoData creation code outlined above. Please help. Thank you!