[BioC] how to handle technical replicates and biological replicates using limma

[Ren Na]

Dr. Gordon Smyth,

Thank you for answering my question.
I still have some questions about how to fit mouse effect for my data, 
My purpose is to find significantly expressed genes between mutant and 
wild type mice. I tried to figure it out by reading "Limma user's guide" 
and came up with the following solution, is it right? or doesn't make 
any sense?
My targets file is:
SlideNumber     FileName        Cy3     Cy5        Target1        Target2
1                      1391.spot        wt1      mu1          wt1      
         mu1
2                      1392.spot        mu1     wt1           mu1      
        wt1
3                      1340.spot        wt2      mu1          wt2      
         mu1
4                      1341.spot        mu1     wt2           mu1      
        wt2
5                      1395.spot        wt3      mu1          wt3      
         mu1
6                      1396.spot        mu1     wt3           mu1      
        wt3
7                      1393.spot        wt1      mu2          wt1      
         mu2
8                      1394.spot        mu2     wt1           mu2      
        wt1
9                      1371.spot        wt2      mu2          wt2      
         mu2
10                    1372.spot        mu2     wt2           mu2         
     wt2
11                    1338.spot        wt3      mu2          wt3         
      mu2
12                    1339.spot        mu2     wt3           mu2         
     wt3
13                    1387.spot        wt1      mu3          wt1         
      mu3
14                    1388.spot        mu3     wt1           mu3         
     wt1
15                    1399.spot        wt2      mu3          wt2         
      mu3
16                    1390.spot        mu3     wt2           mu3         
     wt2
17                    1397.spot        wt3      mu3          wt3         
      mu3
18                    1398.spot        mu3     wt3           mu3         
     wt3
mu1, mu2, and mu3 are different mice which are biological replicates, 
and wt1, wt2 and wt3 are different mice which are biological replicates.
 > targets <- readTargets()
#  I try to include mouse effect in the following way:
 > t1 <- model.matrix(~0+factor(targets$Target1))
 > t2 <- model.matrix(~0+factor(targets$Target2))
 > t3 <- t1+t2
 > design <- cbind(Dye=1, 
MuvsWt=c(1,-1,1,-1,1,-1,1,-1,1,-1,1,-1,1,-1,1,-1,1,-1), t3 )
 > pair <- c(1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9)
 > corfit <- duplicateCorrelation(MA,design,block=pair)
 > fit <- lmFit(MA,design,block=pair,correlation=corfit$consensus)
 > fit <- eBayes(fit)
 
Thanks again for your time.

Ren