[BioC] Illumina beadlevel expression data

[Matt Ritchie]

Hi Ina,

> One other question - to create bead summary data using
> createBeadSummaryData, would you recommend to summarize the unlogged or
> logged bead intensities?
> Thanks again, Ina

We find it useful to summarise the data on the log scale - then the means
and SEs are on the same scale that you would typically do your differential
expression analysis on (assuming you want to use limma, SAM or similar for
DE).

Note that if you want to use vst (variance stabilising transformation) from
lumi, you will need to summarise the data on the original scale before
applying this transformation (I believe it models the mean-variance
relationship on the original scale.  See ?normaliseIllumina in beadarray and
?lumiT and ?lumiN in lumi for more information).

Best wishes,

Matt