[BioC] limma and matrix design for single channel arrays - in search of a primer.

James W. MacDonald jmacdon at med.umich.edu
Tue May 19 15:48:28 CEST 2009


Hi Massimo,

Massimo Pinto wrote:
> I am trying to analyze a microarray experiment that was completed on an
> Agilent single color array platform.
> Using Agi4x44Preprocess, I happily went through the early steps of reading,
> normalization, filtration. Ended up on impressive heatmaps etc.
> 
> Time to move to limma, now. I've been through limma's userguide
> (limma_2.18.0), chapter 7.2 (Affymetrix and other single channel designs)
>  and I am trying to make sense of the construction of a design matrix.  I
> know my basics matrix multiplications, but I don't understand the sense of
> those "-1" in my case of single color arrays. More generally, I had made
> some specifications in my targets file, with regards to what each population
> was, and I am missing how to feed that info into design and contrast
> matrices to operate with limma.
> Is there any reading that you could recommend?

The limma User's Guide is probably the best thing to read, although the 
BioC monograph has some information as well: 
http://www.bioconductor.org/pub/docs/mogr/

However, I don't know what -1s you are talking about. Section 7.2 
doesn't even show the design matrix, and there aren't any -1s in the 
design matrix that is constructed therein:

 > design <- model.matrix(~0 + factor(c(1,1,1,2,2,2,3,3,3)))
 > colnames(design) <- paste("group", 1:3, sep="")
 > design
   group1 group2 group3
1      1      0      0
2      1      0      0
3      1      0      0
4      0      1      0
5      0      1      0
6      0      1      0
7      0      0      1
8      0      0      1
9      0      0      1
attr(,"assign")
[1] 1 1 1
attr(,"contrasts")
attr(,"contrasts")$`factor(c(1, 1, 1, 2, 2, 2, 3, 3, 3))`
[1] "contr.treatment"

And I think using your targets file is covered in some detail as well. 
In your case something like

design <- model.matrix(~0 + targets$Culture)

should get you pretty close to where you want to be.

Best,

Jim


> Thank you in advance.
> Massimo P.
> 
> a sketch of a sample "targets" file follows
> 
>                      X         FileName Treatment GErep Biorep Culture Array
> 41745_1_1_5    t0_1Gy1  41745_1_1_5.txt       1Gy     1      1      t0 41745
> 41844_1_4_6    t0_1Gy2  41844_1_4_6.txt       1Gy     1      2      t0 41844
> 41744_1_1_7    t0_1Gy3  41744_1_1_7.txt       1Gy     1      3      t0 41744
> 41742_1_4_8    t0_1Gy4  41742_1_4_8.txt       1Gy     1      4      t0 41742
> 41745_1_2_13 6moA_1Gy1 41745_1_2_13.txt       1Gy     2      1    6moA 41745
> 41742_1_3_14 6moA_1Gy2 41742_1_3_14.txt       1Gy     2      2    6moA 41742
> 41741_1_3_15 6moA_1Gy3 41741_1_3_15.txt       1Gy     2      3    6moA 41741
> 41844_1_3_16 6moA_1Gy4 41844_1_3_16.txt       1Gy     2      4    6moA 41844
> 41745_1_3_21 6moC_1Gy1 41745_1_3_21.txt       1Gy     3      1    6moC 41745
> 41844_1_2_22 6moC_1Gy2 41844_1_2_22.txt       1Gy     3      2    6moC 41844
> 41744_1_2_23 6moC_1Gy3 41744_1_2_23.txt       1Gy     3      3    6moC 41744
> 41743_1_3_24 6moC_1Gy4 41743_1_3_24.txt       1Gy     3      4    6moC 41743
> 

-- 
James W. MacDonald, M.S.
Biostatistician
Douglas Lab
University of Michigan
Department of Human Genetics
5912 Buhl
1241 E. Catherine St.
Ann Arbor MI 48109-5618
734-615-7826



More information about the Bioconductor mailing list