[BioC] Limma and randomized block design for time-course microarray experiment

[Gordon Smyth]

Dear Serge,

That's what I do. For example, the time course analysis in

Peart, M. J., Smyth, G. K., van Laar, R. K., Richon, V. M., Holloway, 
A. J., Johnstone, R. W. (2005). Identification and functional 
significance of genes regulated by structurally diverse histone 
deacetylase inhibitors.  Proceedings of the National Academy of 
Sciences of the United States of America.102, 3697-3702.

was done that way. Each time course was hybed at one time, as in your 
experiment.

Please note, this doesn't constitute a detailed check of your code, 
just a confirmation of the idea of using blocks for a time courses 
done at one time.

Best wishes
Gordon

>Date: Mon, 23 Jul 2007 15:50:47 +0200
>From: "Serge Eifes" <serge.eifes at lbmcc.lu>
>Subject: [BioC] Limma and randomized block design for time-course
>         microarray      experiment
>To: <bioconductor at stat.math.ethz.ch>
>Cc: serge.eifes at lbmcc.lu
>
>Dear all,
>
>We have performed a time-series experiment (2h, 6h, 10h, 48h,  ?) on
>dual-channel arrays where we want to compare gene expression between treated
>and time-matched untreated cells. For every time-point three hybridizations
>were performed including one dye-swap. Each time-course was hybridized at
>the same time.
>
>So for a given time point we have 3 different microarrays comparing:
>Tx -> Ux
>Tx ->Ux
>Ux ->Tx
>
>T=treated cells?; U=untreated cells at corresponding time-point?; x=given
>time-point.
>
>Is it possible in this case (unbalanced design) to use randomized blocks in
>LIMMA to estimate variability between the three microarray batches
>(time-courses)?
>
>If affirmative, the R code I would use looks as follows:
>
> > design = cbind( T2vsU2=c(1,1,-1,0,0,0,0,0,0,0,0,0),
> > T6vsU6=c(0,0,0,1,1,-1,0,0,0,0,0,0),
> > T10vsU10=c(0,0,0,0,0,0,1,1,-1,0,0,0),
> > T48vsU48=c(0,0,0,0,0,0,0,0,-0,1,1,-1))
> > blocks = c( 1,2,3,1,2,3,1,2,3,1,2,3)
> > dupcor = duplicateCorrelation(MA, design=design, block=blocks)
>
>Is this piece of code appropriate to model the randomized blocks (the three
>time-courses)?
>
>
>Many thanks in advance,
>
>Serge
>
>
>Serge Eifes
>Laboratoire de Biologie Moleculaire et Cellulaire du Cancer (LBMCC)
>Hopital Kirchberg
>9,rue Edward steichen
>L-2540 LUXEMBOURG