[BioC] Defining and handling replicates

[Jonathan Arthur]

Hello,

I am using Affymetrix arrays to compare two groups of samples (5 in one 
group and 4 in the other). I have been using affy, affylmGUI, and limma 
for the analysis. Some of the samples in the latter group have 
replicates and I'd like to know how to handle these. My questions are:

1) In one case, the replicate is drawn from the same sample but at a 
later time point. Is this a technical replicate (because it is the same 
sample but different chip), biological replicate (i.e., the different 
time point make it *effectively* a different sample), or neither (and 
thus can't be used)?

2) I assume I need to average the gene expression of the technical 
replicates before doing the analysis and treatment them as *one* sample? 
Is this correct?

3) Assuming I do need to average the expression of the technical 
replicates, are there methods in affy or limma to do this? Or do I need 
to do it manually with something like:

Data <- ReadAffy()
eset <- rma(Data)
# do something here to create a new eset where the technical replicate 
columns have been replaced by a single column averaging the two
# go on to limma analysis

Thanks,

Jonathan

-- 
Dr Jonathan Arthur
Sesqui Lecturer in Bioinformatics
Central Clinical School, Faculty of Medicine and SUBIT
Medical Foundation Building, K25
University of Sydney
Ph: +61 2 9036 3132
Email: jarthur at med.usyd.edu.au