[BioC] help with design matrix

[Gordon Smyth]

You need to define a contrast matrix. See section 7.3 of the limma User's 
Guide.

Gordon

>Hello Everyone,
>I am trying to compare two red-labeled samples via a green-labeled
>reference RNA. Here is an example of a script I've run:
>
>  > library(limma)
>  > files <- dir(pattern="*.gpr")
>  > RG <- read.maimages(files, source="genepix")
>Read 099.gpr
>Read 101.gpr
>Read 105.gpr
>Read 107.gpr
>Read 111.gpr
>Read 113.gpr
>  >
>  > targets <- readTargets()
>  > targets
>   FileName Cy3 Cy5 Name
>1  099.gpr Ref  A3   99
>2  101.gpr Ref  A5  101
>3  105.gpr Ref  A3  105
>4  107.gpr Ref  A5  107
>5  111.gpr Ref  A3  111
>6  113.gpr Ref  A5  113
>  >
>  > RG$genes <- readGAL()
>  > design <- designMatrix(targets,ref="Ref")
>  > design
>   A3 A5
>1  1  0
>2  0  1
>3  1  0
>4  0  1
>5  1  0
>6  0  1
>
>My intention is to compare "A3" with "A5" via a common reference, "Ref",
>similar (I think) to the example in the limma User's guide section 8.2
>"ApoAI Knockout Data ...". What am I doing wrong? Is my Targets.txt file
>set up incorrectly? In its current state it results in topTable giving
>me a list that seems appropriate for contrast between Cy3 and Cy5. I've
>also tried creating a "design" object like the one in the example
>manually by writing it in a text editor and importing it with
>read.table(). I just get nonsense back from that.
>
>I'm running BioC v1.3 R 1.8.1 on linux kernel 2.6.4
>
>Thanks in advance for the help.
>
>-Dennis
>