I have a experiment like this:

Treated:

T_Sampe_A1, T_Sample_A2 ,T_Sample_A3, T_Sample_B1, T_Sample_B2, T_Sample_B3

Untreated:

U_Sampe_C1,U_Sample_C2, U_Sample_C3, U_Sample_D1, U_Sample_D2, U_Sample_D3

Here I would expect some DE genes between A vs B samples and C vs D samples but I just want to compare Treated Vs Untreated. I would like to know which one would be the best way to get DE genes between treated vs untreated:

1. Just create a design like c(rep("t",6), rep("u",6)) and perform differential analysis ?

2. create a design for glm like:

cond= c(rep("t",6), rep("u",6))
internal=c(rep("t_a",3),rep("t_b",3),rep("u_c",3),rep("u_d",3))
design = model.matrix(~internal+cond)
y <- estimateGLMCommonDisp(y, design, verbose=TRUE)
y <- estimateGLMTrendedDisp(y, design)
y <- estimateGLMTagwiseDisp(y, design)
fit <- glmFit(y, design)
lrt <- glmLRT(fit)​

and use design to compare treated vs untreated ? Will it make any difference ?