Hi

I am trying to get a camera test working after my differential expression in limma. I do not know how to set up camera properly here, because contrast 1,2,3 is not the same as co efficients 1,2,3 in my DE analysis - because the results are not right. I am stuck, any words of wisdom?

colnames(design) <- c('class1', 'class2', 'class3', 'age', 'gender')
matrix <- data.matrix(data3)
fit <- lmFit(matrix, design)
contrast.matrix <- makeContrasts(class1-class2, class3-class2, class3-class1, levels=design)
fit2 <- contrasts.fit(fit, contrast.matrix)
fit2 <- eBayes(fit2)
anova <- topTableF(fit2, number=Inf)

top2 <- topTable(fit2,coef=3,number=Inf,sort.by="P") # I choose my comparisons here

# parse .gmt files

setwd("~/Desktop/gmt_files")
x <- scan("c2.cp.kegg.v5.0.symbols.gmt", what="", sep="\n")
y <- strsplit(x, "[[:space:]]+")
names(y) <- sapply(y, `[[`, 1)
y <- lapply(y, `[`, -1)

c2.indices <- ids2indices(y, row.names(data3))
results <- camera(data3, c2.indices, design,contrast=3) # this is the bit I am not sure about

Thanks,

Chris

Presumably you want to run CAMERA on the third comparison in contrast.matrix. If so, then you can do:

results <- camera(data3, c2.indices, design, contrast=contrast.matrix[,3])

This will use the DE between classes 1 and 3, same as the DE identified in top2.