High-Throughput Strand-Specific mRNA Library Preparation for Illumina Sequencing from Total RNA Isolated from Normal and Cancerous Ovary Tissue

The cost benefits of whole genome and transcriptome next generation sequencing (NGS) have revolutionized genetic analysis. Determination of regulatory changes in cancer cells by whole transcriptome analysis has proven useful for tailoring treatment options for patients. An efficient, consistent and reliable whole transcriptome library preparation method is necessary for successful transition of the NGS technologies from research to clinical applications. Due to the labile nature of mRNA and the tedium of processing large numbers of patient samples, automation would greatly improve the reliability and throughput of whole transcriptome library preparation. We have explored the automation of whole transcriptome library preparation from total RNA of normal and cancerous ovary tissue. The Apollo 324™ System (IntegenX Inc.) was used to isolate poly A mRNA from 6 to 48 samples of total RNA and prepare strand-specific mRNA libraries for sequencing on the Genome Analyzer IIx (Illumina). The cDNA output was amplified on a bench thermocycler to yield whole transcriptome libraries in a single eight hour day. While most conventional RNA-Seq library preparation methods convert mRNA to cDNA, in our strand specific library preparation, we ligated the adapters directly to fragmented mRNA to preserve strand polarity. Preserving strand polarity of the transcript reduces the bioinformatics bottleneck. A commercially available manual strand-specific library was used as the bench control for comparison and validation of our library preparations. The gene expression profile of 12 up-regulated and down-regulated genes was equivalent between the high-throughput libraries and the published data. The isolated polyA mRNA had an average of 0.3% rRNA contamination from 500 ng of total RNA. There was a 75% time reduction for automated library preparation from total RNA compared with conventional library preparation methods. There was 90% correlation of gene expression between the automated and bench library preparation methods for up to 48 samples.