Dear all,
I need some help on matrix design and B statistics by using limma package.
I want to compare gene expression in 2 groups of cDNA samples.
The experiment compares 4 treated mice(#1,#2,#3,#4) and 4 control mice
(#5,#6,#7,#8).
The target file is
FileName Cy3 Cy5
mice1.spot Control_#5 Treat_#1
mice2.spot Treat_#1 Control_#5
mice3.spot Control_#6 Treat_#2
mice4.spot Treat_#3 Control_#7
mice5.spot Control_#8 Treat_#4
The first slide (mice1.spot) and the second slide(mice2.spot) are
dye-swap. There is no common reference. There are 3 replicated spots of
each gene on each array (384 genes in total).
MA is an object of class marrayNorm, below is what I did.
>design <- c(1,-1,1,-1,1)
>cor <- duplicateCorrelation(MA,design,ndups=3)
>cor$consensus.correlation
[1] 0.506
>fit <- lmFit(MA,design,ndups=3,correlation=cor$consensus.correlation)
>fit <- eBayes(fit)
>topTable(fit,n=20,adjust="fdr")
The result is,
ID M A t P.Value B
348 -1.3 10.8 -3.98 0.577 -4.47
371 -1.91 11.5 -3.36 0.577 -4.47
172 -2.56 13.4 -3.36 0.577 -4.47
273 -0.98 10.3 -3.22 0.577 -4.48
...
It seems this is no evidence of differential expression. But if I use the
first three slides to do analysis, design <- c(1,-1,1),the result is good,
B>5, P.Value is very small. I am wondering if my design matrix is right?
Many thanks in advance and best regards.
Michelle
______________________________________________
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
