Trying to set up a design matrix for two colour microarry exp. Two treatments A vs B. 10 participant.  Each participant has a control/baseline sample of muscle hyb'd with a sample of muscle with treatments A and B.  Targets

Filename     cy3     cy5

array1           r1     A

array2           r1     B

array3            r2     A

array4            r2     B

array 5           r3     A

array 6           r3     B

array 7           r4    A

array 8           r4    B

etc up to the 10th patient.  Would the matrix be 

Group <-factor(c("A","B","A","B","A","B","A","B","A","B","A","B","A","B","A","B"), levels=c("A","B"))

>design <- cbind(A=c(1,0,2,0,3,0,4,0,5,0,6,0,7,0,8,0,9,0,10), B=c(0,1,0,2,0,3,0,4,0,5,0,6,0,7,0,8,0,9,0,10))

>fit <- lmFit(MA, design)

Thanks James

No. You have accounted for the blocked design by using a control muscle from each subject. In other words, for each array, the log ratio is log(treatment/control), which measures the difference in treatment after controlling for the baseline control levels.

Since the design of your experiment already accounts for the intra-block variability, I would use a cell means model.

design <- model.matrix(~ 0 + Group)

If you just care about treatment A vs control and treatment B vs control, then you would do

fit <- lmFit(MA, design)

fit2 <- eBayes(fit)

trtA <- toptable(fit2,1)

trtB <- topTable(fit2,2)

But if you also want to know about differences between treatments, you need a contrasts matrix

contrast <- matrix(c(1,0,0,1,1,-1), ncol = 2)

fit <- lmFit(MA, design)

fit2 <- contrasts.fit(fit, contrast)

fit2 <- eBayes(fit2)

trtA <- toptable(fit2,1)

trtB <- topTable(fit2,2)

trtAvB <- toptable(fit2,3)