Background Characterization of whole transcriptomes from small RNA samples (less than 1 ng total RNA) has been a major challenge. Typically, samples are extracted and purified, amplified by in vitro transcription (IVT), and then subjected to genome-scale analysis. Using spike-in controls designed to mimic mammalian mRNA species, we assessed whether the linear amplification process was faithfully reproducing, or introducing bias into, the samples mRNA profile. We focused especially on the confidence of the transcript level measurement to highlight its role in differential gene expression analyses. Results IVT reproduces gene expression profiles down to approximately 100 absolute input copies. Repeat amplification experiments show high concordance, with noise increasing inversely with copy number. The amplification appears biased towards species with more copies, making the IVT reaction somewhat non-linear under low total RNA input/long IVT conditions. Conclusions First round linear amplification nearly preserves the gene expression information within a sample down to the 100 copy level, regardless of total input sample amount. The non-linearity of the IVT reaction will likely not cause major challenges for most gene expression studies, though some transcript-specific differential gene expression values may need adjustment for amplification bias. The variability of the amplification increases predictably with decreasing input copy number. Hence, properly executed IVT reactions are unlikely to significantly bias gene expression profiles, and proper experimental design with reasonable detection limits is likely to yield differential gene expression capability even between small copy number transcripts.
Citation: Nucleic Acids Research
Pub Type: Journals
in vitro transcription, ERCC, mRNA, aRNA