I have a microarray dataset with and I have labeled the data such that it looks something like this:

[9967] piRNA          piRNA          piRNA          piRNA          piRNA          piRNA         
[9973] piRNA          piRNA          piRNA          piRNA          piRNA          piRNA         
[9979] snoRNA         snoRNA         snoRNA         snoRNA         snoRNA         snoRNA         
[9985] tiRNA          tiRNA          tiRNA          tiRNA          tiRNA          tiRNA         
[9991] miRNA          miRNA          miRNA          miRNA          miRNA          miRNA              

So basically I have a bunch of different probe-types in a single-channel micro-array and I want to make a contrast to see which are DE.

I have tried to do this as follows:

design<-factor(levels(probe.type1))
design<-model.matrix(~0+design)
dim(design)
dim(E.ncRNA1)
subE.ncRNA1<-E.ncRNA1[ ,1:12]
fit<-lmFit(subE.ncRNA1, design)
contrast.matrix <- makeContrasts(piRNA-miRNA, levels=design)

trying <- contrasts.fit(fit, contrast.matrix)
try <- eBayes(trying)
trying<-eBayes(trying)
volcanoplot(trying)

The problem occurs where I try to make the contrasts:

contrast.matrix <- makeContrasts(pivsmi=piRNA-miRNA, levels=design)
Error in makeContrasts(pivsmi = piRNA - miRNA, levels = design) : 
The levels must by syntactically valid names in R, see help(make.names).  Non valid names: designputative miRNA,designsub scRNA,designvault RNA

Can anyone tell me what I'm doing wrong? I plan to submit this Q to the BioConductor help forum as well.

Hi Christopher,

Don't take this the wrong way, but from this question (of what I can parse of it) and your posting history, it seems like you would benefit from talking with a local statistician, or a more senior student, or your mentor, etc. to get a better idea of what types of questions you can ask (and how to ask them) using a linear model framework.

It doesn't seem like the contrast you are trying to make makes much sense (ie. piRNA-miRNA) unless you have two different conditions encoded as such. From your first code block, however it looks like you are trying to name the rows of your matrix as such?

Which also doesn't make much sense ... which is why I suggest you sit down with someone who's worked through some of this before and get some 1:1 time with them so that they can help you work through the basics.