[BioC] Using a weight for Illumina data? (limma)

Benno Puetz puetz at mpipsykl.mpg.de
Tue Jan 22 14:20:56 CET 2008

As I have a lot of genes which show essentially no expression, I thought
of using the Detection (or rather 1-detection) as weight in lmFit. As I
have not found any mention of this yet, I wanted to ask for comments
whether that would be reasonable.

When subsequently using eBayes(), this would probably reduce the 
influence of the low variance in those genes?

Benno Pütz
Statistical Genetics
Max-Planck-Institut of Psychiatry             Tel.: +49-89-30622-222
Kraepelinstr. 10                              Fax : +49-89-30622-601
80804 Munich, Germany

More information about the Bioconductor mailing list