Isoform Sequencing: Unveiling the Complex Landscape in Eukaryotic Transcriptome on the PacBio® RS II

Advances in RNA sequencing have accelerated our understanding of the transcriptome, however isoform discovery remains challenging due to short readlengths.  The Iso-Seq™ method provides a new alternative to sequence full-length cDNA libraries using long reads from the PacBio RS II.

Identification of long and often rare isoforms is demonstrated with rat heart and lung RNA prepared using the Clontech^® SMARTer^® cDNA preparation kit, followed by agarose-gel size selection in fractions of 1-2 kb, 2-3 kb and 3-6 kb. For each tissue, 1.8 and 1.2 million reads were obtained from 32 and 26 SMRT^® Cells, respectively. Filtering for reads with both adapters and polyA tail signals yielded >50% putative full-length transcripts. To improve consensus accuracy, we developed an isoform-level clustering algorithm ICE (Iterative Clustering for Error Correction), and polished full-length consensus sequences from ICE using Quiver. This method generated full-length transcripts up to 4.5 kb with ≥ 99% post-correction accuracy. Compared with known rat genes, Iso-Seq method not only recovered the majority of currently annotated isoforms, but also several unannotated novel isoforms with identified homologs in the RefSeq database. Additionally, alternative stop sites, extended UTRs, and retained introns were detected.

To conclude, several advantages over short-read technologies were demonstrated using the Iso-Seq method: (1) full-length transcript sequencing up to 6 kb, and possibly longer; (2) no assembly required; (3) no reference genome required. We believe high quality sets of transcripts identified by Iso-Seq method are critical for understanding transcriptome complexity and discovering novel biological insights.