We would like to carry out an experiment on a cell line treated with cytoprotective agents. Given state of technology and analysis algorithms (normalization, differential gene expression analysis) what is a reasonable level of biological replication? I noticed that many paper cite 3 arrays per condition, which suprised me, I would expect higher numbers. But I would not object to lower numbers if it would be seen as quite standard :).
-- I am aware that biological variability is key factor.
Also Illumina offer 2 slides by 12 chips. Is it reasonable to make cross-slide comparisons?
thanks for guidance