The question is:

We will run the analysis with two factors but without an interaction term. You need to be careful in how you specify the design, as results() will by default return the results for the variable that is specified last. Of course, you can also explicitly specify the contrast you want to perform.

my code:

res<-results(dds)
dds$condition <- factor(dds$condition, levels = c("untreated","treated"))
UvsT<-results(dds, contrast=c('condition', 'untreated', 'treated' ))
UvsT
plot(res$log2FoldChange, UvsT$log2FoldChange, xlab='first analysis', ylab='second analysis')

the main question is :

How high is the correlation (e.g. Pearson correlation coefficient) between the log2FoldChange for untreated vs treated in the unifactorial and the two-factorial analysis?

Can anybody help me to calculate the correlation?

Many thanks 

Hira

The correlation will probably be high, but the important thing is to choose a design that makes sense for your samples and stick with it.

Do you want to describe your experiment and we can help to suggest an appropriate design?