We considered an association as significant, if it was found at least in four microarray experiments (see below).
First, only a single microarray experiment was considered.
In a two-color microarray experiment, we consider the issues of determination of which mRNA samples are to be labeled with which fluorescent dye and which mRNA samples are to be hybridized together on the same slide.
A potential pitfall of this strategy, however, is whether independently generated datasets are combinable since the systematic bias between microarray experiments is commonly considered substantial.
This comparison was done 'qualitatively'- looking for gene products which showed homodirectional changes (i.e. changed in the same direction) in our proteomic dataset and in microarray experiments, but not considering the magnitude of these changes in these different datasets.
The composition of the EST set used in these microarray experiments has previously been considered specifically with respect to the complement of developmental signalling pathway components [ 2, 3], but this paper is the first to examine broad scale changes in gene expression during development for any cnidarian.
Of the 3.7 Mb of transcribed sequence found only by modENCODE, 1.3 Mb is intronic to our models and much of the remaining 2.4 Mb likely reflects poly(A)− transcripts detected by modENCODE total RNA and microarray experiments that were not considered in our analyses.
Furthermore, posttranslational mechanisms such as phosphorylation, important for the biological activity of PTEN, are not considered in microarray experiments.
Thus, in the light of all other well known technologically caused sources of error in microarray experiments which have to be considered carefully when the experiments are designed, the error introduced by RNA amplification is small.
In these sections, the authors evaluate the different microarray platforms, probe labeling, image analysis, various crucial and essential quality controls, etc. Owing to the large amount of data generated by microarray experiments, statistical analysis should be considered as the cornerstone of such studies.
