[BioC] LIMMA - generation of design vector for lm using duplicate spots?

James Wettenhall wettenhall at wehi.edu.au
Wed Jun 23 14:40:33 CEST 2004

On Wed, 23 Jun 2004, Jakob Hedegaard wrote:
> When ignoring duplicate spots, the functions "readTargets", 
> "modelMatrix" and "makeContrasts" helps to construct the 
> design matrix.
> Are there similar functions for "automatic" generation of the 
> design vector when considering duplicate spots using 
> "duplicateCorrelation" and "gls.series"?


In limma, the within-array duplicate spots are handled in a very 
different way from the replicate arrays, so while the design matrix 
depends on the replicate array structure, it doesn't depend on 
the within-array duplicate spot structure, so you can just use 
the same functions: readTargets, modelMatrix etc.

Instead of using a design matrix to tell limma about your 
duplicate spot structure, you use the ndups and spacing 
arguments of the duplicateCorrelation function.

 Hope this help,

More information about the Bioconductor mailing list