Abstract
Circular RNA (circRNA) is a class of covalently joined non-coding RNAs with functional roles in a wide variety of cellular processes. Their composition shows extensive overlap with exons found in linear mRNAs making it difficult to delineate their composition using short-read RNA sequencing, particularly for long and multi-exonic circRNAs. Here, we use long-read nanopore sequencing of nicked circRNAs (circNick-LRS) and characterize a total of 18,266 and 39,623 circRNAs in human and mouse brain, respectively. We further develop an approach for targeted long-read sequencing of a panel of circRNAs (circPanel-LRS), eliminating the need for prior circRNA enrichment and find >30 circRNA isoforms on average per targeted locus. Our data show that circRNAs exhibit a large number of splicing events such as novel exons, intron retention and microexons that preferentially occur in circRNAs. We propose that altered exon usage in circRNAs may reflect resistance to nonsense-mediated decay in the absence of translation.
Highlights
Circular RNA is a class of covalently joined non-coding RNAs with functional roles in a wide variety of cellular processes
The circRNAenriched pool was nicked by gentle hydrolysis prior to long-read sequencing, followed by 3′ end polyadenylation to align with the standard Oxford Nanopore Technology (ONT) protocol (Fig. 1b)
We apply ONT longread sequencing technology to describe the exon composition of full-length circRNAs as well as the corresponding mRNA. This approach circumvents the limitation of second-generation pairedend sequencing methods where the read length is limited to ~200 nucleotides adjacent to the back-splice junction (BSJ), insufficient to read through most circRNAs
Summary
Circular RNA (circRNA) is a class of covalently joined non-coding RNAs with functional roles in a wide variety of cellular processes Their composition shows extensive overlap with exons found in linear mRNAs making it difficult to delineate their composition using short-read RNA sequencing, for long and multi-exonic circRNAs. Here, we use long-read nanopore sequencing of nicked circRNAs (circNick-LRS) and characterize a total of 18,266 and 39,623 circRNAs in human and mouse brain, respectively. ONT has been used to sequence poly (A)-purified RNA and reported a significant number of transspliced RNAs24; the inclusion of a poly(A) selection step prevented the detection of circRNAs. Another study combined nanopore sequencing with a PCR-based approach by using endto-end divergent primers to create BSJ reads and find different isoforms of circNPM1; this study was limited to this single circRNA25. It raises the possibility that circRNA-specific RNA elements may exert novel functions which are absent in the linear transcriptome
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
