getting differential expressed genes between tissues, species and time series

Question

[zizigolu]

Hi,

I have a microarray experiment (GSE63706) with 6 time points, tissues, three species and two replication. I am going retrieve genes with fold change more than two.

May someone tell me a design for limma because I got confused as already.

I have a design for treatment and normal in fout time points but I can adapt for my case

The design is like below

X <- ns(targets$time.h, df= 2)

Group <- factor(targets$treatment)
design <- model.matrix(~Group*X)
colnames(design) <- c("Intercept", "Groupheat","Control_X1", "Control_X2","Heated_X1", "Heated_X2")
design
fit <- lmFit(data, design)
fit <- eBayes(fit)

Answer 1

[andrew.j.skelton73]

You'll need to make a contrast matrix to design the actual comparisons you make. Check out the limma users guide on making contrast matrices. I'd also very strongly suggest that you read up on interaction models, if you're not from a statistical background then it can get very confusing, very quickly. Check out section 9.5.4 of the limma users guide, which gives an example of classic interaction models.

[zizigolu]

Thank you, i was reading that is why i got totally confused