Description
To address the neglected possibility for global mRNA changes in microarray experiments we developed a simple method to generate external controls for Affymetrix microarrays to allow these platforms to measure absolute mRNA expression at the global level. We used publicly available data to select probesets that never detect endogenous transcripts, and used PCR and IVT to generate synthetic mRNAs corresponding to them. After quality control and testing, these control transcripts were spiked into total RNA samples from plants before and after 24 h of cold treatment. Due to changes in the proportion of mRNA, these data reveal intensity-dependent bias in expression estimates based on standard all-gene normalizations. When not accounted for, this leads to false classification of the differential expression for thousands of genes.