A set of 11 genes was tested by RT-qPCR using the same samples used in the microarray experiments to validate the microarray platform results.
Our results show that experimental conditions, laboratory where the experiments were performed and different microarray platforms can result in significant variability in gene expression patterns from similar sources of cells.
Though gene expression of tomato tissues has been widely investigated in the years, thanks to both EST sequencing and different microarray platforms, the resulting collections are heterogeneous in terms of experimental approaches employed, genotypes investigated and conditions considered.
Although some gene classifiers identified in our study were common to other reported GEP studies, the absence of some previously identified key classifiers may be due to variable probe make-up across different microarray platforms or resulting from differences in the type of diagnostic classes used in our classification compared to most published GEP studies [ 21].
However a similar parametric analysis of data from a distinct microarray platform also results in the identification of a set of differentially expressed genes that dissimilar to a degree that is beyond what would be expected from false-positive errors alone.
In summary, this study objectively evaluates several published signatures in independent cohorts from diverse microarray platforms and unifies results of previous gene expression studies in breast cancer.
Since 2004, the vast majority (29/32) of technical papers comparing microarray platforms have generated results that show a moderate to high degree of correlation among the technologies.
Xu et al. [ 26] compared transcriptome profiles that were generated by Illumina RNA-Seq and Affymetrix microarray platforms, and their results suggested that RNA-Seq is more advantageous for detecting genes with low-level expression compared with microarrays.
In this study we have excluded discrepancies due to different microarray platforms and our results argue against recent criticism that gene expression profiling may not be robust enough to be useful for clinical application [ 44, 45].
While these analyses emphasized critical issues such as the compatibility across different microarray platforms, they tended to result in conflicting conclusions because the "relative to relative" nature of such approaches.
