UC San Diego

Quantification of global transcripts expression is a key step towards developing system-level understanding in biology. Probe independent RNA-seq provides digital estimation of transcript abundance with dynamic range large enough to accurately quantify the majority of complex mammalian transcriptomes. However, a reliable quantification of low abundant transcripts from limited amounts of mRNA has remained a challenge for RNA-seq. The widely used RNA-seq protocol requires 1-10 ng of mRNA to generate robust sequencing libraries restricting its application in disciplines where obtaining such amounts of mRNA is challenging, such as in developmental biology, stem cell biology and forensics. To address this issue, we developed a novel RNA-seq methodology (DP-seq) that uses a defined set of 44 heptamer primers to amplify majority of the mammalian transcripts from limiting amounts of mRNA, while preserving their relative abundance. DP-seq reproducibly yields high levels of amplification from as low as 50 pg of mRNA (50-100 mammalian cells) with a dynamic range of over five orders of magnitude in RNA concentrations. A novel two-step amplification step utilizing a combination of mesophilic and thermophilic polymerases was devised to achieve efficient amplification from the heptamer primers. Furthermore, we exploited PCR biases observed in our methodology to reduce the representation of highly expressed ribosomal transcripts by more than 70% in our sequencing libraries. We validated DP-seq on lineage segregation model in early stem cell cultures achieved by modulating TGF[Beta] pathway. DP-seq accurately quantified the majority of the low expressed transcripts and revealed novel lineage markers and putative TGF[Beta] target genes. Similarly, by using DP- seq we functionally characterized dedifferentiated neurons and astrocytes and found the cell cycle, Wnt signaling and the focal adhesion pathways to be involved in the maintenance of their undifferentiated state. Finally, we compared DP-seq with other amplification-based strategies and found similar transcriptome coverage and overlapping technical noise. Interestingly, the technical noise increased significantly when ultra-low amount of mRNA (single cell level) was used, irrespectively of the methodology. In conclusion, this study provides an economical and efficient solution for sequencing library generation using low amounts of mRNA thereby increasing the applicability of RNA-seq to a wider spectrum of biological systems