Computational protocol: An intriguing RNA species—perspectives of circularized RNA

Similar protocols

Protocol publication

[…] Due to its circular structure, the migration rate of circRNA is different from that of linear, interlocked, and lariat RNAs in eletrophoresis (Hansen et al., ; Tabak et al., ). CircRNA can also be characterized through Gel trapping since it can attach to a well when mixed with melted agarose due to its circular structure (Schindler et al., ). The more credible way to recognize circRNA for a specific gene is using PCR with out-facing primer pair (located near the back-splicing sites on the corresponding linear transcript), though false-positive cannot be excluded because of various reasons. Both genome rearrangement and trans-splicing can produce exon shuffling transcripts, and dislocated exons produced by template switch in in vitro reverse transcription (RT) is another artifact in circRNA detection using RT-PCR. Therefore, more accurate methods are required to confirm the positive results, such as using the same out-facing primer pair for the genomic DNA as a control when PCR was carried out. There are optional ways to improve the accuracy in identifying circRNA, such as enriching circRNA by degrading linear RNA with exonuclease—RNase R or tobacco acid pyrophosphatase combined with 5′-phosphate-dependent exonuclease (Hansen et al., b). Northern blotting and RNase protection assay can also serve as complementary methods for validating circRNAs.The extremely low abundance of circRNA is a constraint for its PCR-based detection. High-throughput sequencing combined with bioinformatic analysis has greatly facilitated the discovery of this kind of RNA molecule (Table ). Unlike the normal RNA sequencing library preparation step, disposing samples through methods such as ribosomal RNA depletion and poly (A) minus RNA selection alone or in combination with RNase R treatment is indispensable prior to library construction, which has favored the discovery of circRNAs (Danan et al., ; Jeck et al., ; Zhang et al., ).To exclude confounding interferences and obtain convincing results, various algorithms have been developed to analyze high-throughput sequencing data. On the basis of the sequence order or any abnormal exon–exon junction boundaries from annotated exon, scrambled RNA molecules have been distinguished from linear molecules (Memczak et al., ; Salzman et al., ). Reads aligned to the known spliced junctions are considered as linear splicing and thus are filtered out, while reads that only span the 3′ end of the downstream exon and 5′ end of the upstream exon are considered potential signals of circRNA, in which the junction is AG/GU but not the canonical GU/AG (Memczak et al., ). In addition, other innovative strategy or bioinformatic analysis methods can also be applied to identify novel circRNAs (Salzman et al., ).Lastly, to detect low-abundance circRNAs, RNase R nucleases were used to enrich circRNA by eliminating linear RNA prior to RNA-seq library construction (Jeck et al., ; Zhang et al., ). Mapping algorithms dependent (e.g., MapSplice (Jeck et al., ; Wang et al., )) or independent on gene annotation (e.g., TopHat-Fusion (Zhang et al., )) could be used to detect the potential junctions of circRNAs. Experimental validation was followed to confirm the existence of identified circRNAs. […]

Pipeline specifications

Software tools MapSplice, TopHat-Fusion
Application RNA-seq analysis