Abstract
Circular RNA (circRNA) is a poorly understood class of non-coding RNAs, some of which have been shown to be functional important for cell proliferation and development. CircRNAs mainly derive from back splicing events of coding mRNAs, making it difficult to distinguish the internal exon composition of circRNA from the linearly spliced mRNA. To examine the global exon composition of circRNAs, we performed long-read sequencing of single molecules using nanopore technology for human and mouse brain-derived RNA. By applying an optimized circRNA enrichment protocol prior to sequencing, we were able to detect 7,834 and 10,975 circRNAs in human and mouse brain, respectively, of which 2,945 and 7,052 are not currently found in circBase. Alternative splicing was more prevalent in circRNAs than in linear spliced transcripts, and notably >200 not previously annotated exons were used in circRNAs. This suggests that properties associated with circRNA- specific features, e.g. the unusual back-splicing step during biogenesis, increased stability and /or their lack of translation, alter the general exon usage at steady state. We conclude that the nanopore sequencing technology provides a fast and reliable method to map the specific exon composition of circRNA.
