2 protocols
for processed data, please refer to the original publication
Total RNA was extracted using the Trizol (Invitrogen) procedure or RNAeasy/RNAeasy Lipid/miRNeasy (Qiagen) column purification kits as indicated. RNA quality was assessed using an Agilent 2100 Bioanalyzer. Sequencing libraries were prepared using the mRNA-Seq Sample Prep Kit (Illumina) according to the manufacturer's instructions. Briefly, polyadenylated RNA was isolated using a poly-dT bead procedure and then chemically fragmented and randomly primed for reverse transcription. After second-strand synthesis, the ends of the double-stranded cDNA were repaired. After 3'-end adenylation of these products, Illumina Paired-End Sequencing adapters were ligated to the blunt ends of the cDNA fragments. Ligated products were run on gels; 250-300 bp fragments were excised and then PCR-amplified (15 cycles). After column-purification, qualities of the resulting libraries were assessed using Agilent 2100 Bioanalyzers. Potential influences on RNA sequencing results due to different experimenters preparing the libraries were ruled out on the basis of RNA- Seq data comparisons of replicate libraries prepared by the different experimenters (i.e., expression levels derived from replicate libraries were highly correlated: rho > 0.99). The RNA-Seq libraries were each sequenced with 76 (101) cycles in at least one lane of the Illumina Genome Analyzer IIx platform according to the manufacturer's specifications. Technical replicates (i.e., sequencing the same library on different machines) were performed to rule out potential biases during the sequencing step (i.e., expression levels between technical replicates were highly correlated: rho > 0.98). After sequencing, we processed the fluorophore intensity files with the Ibis base caller (version 1.11), in addition to applying the standard Illumina base calling algorithms. As illustrated in Supplementary Note Table 1 and Supplementary Note Figure 1 for a small subset of samples, we found that Ibis significantly increased the number of usable reads and drastically reduced the error rate. This improvement was more pronounced for the sequencing runs processed with early versions of the Illumina pipeline, but remained noticeable even for the latest Illumina release (GA pipeline 1.60). All subsequent analyses were performed on the Ibis-called reads.