What is a circular RNA and how is it different from linear RNA?

A circular RNA (circRNA) is a covalently closed RNA molecule formed by back-splicing. Unlike linear RNAs, circRNAs lack 5' and 3' ends, making them resistant to exonuclease degradation and RNase R digestion.

Why is RNase R treatment important for circRNA-seq?

RNase R selectively degrades linear RNA while leaving circular RNAs intact, enriching the circRNA fraction 10-50 fold and dramatically increasing back-splice junction read detection.

How do you distinguish genuine circRNAs from false positives?

We use multi-algorithm consensus detection (CIRI2, CIRCexplorer2, find_circ), minimum read support thresholds, alignment artefact filtering, cross-sample reproducibility assessment, and experimental validation by RNase R resistance and divergent primer RT-PCR.

What sequencing depth is recommended for circRNA-seq?

Standard: 80-100 million reads per sample without RNase R treatment; 40-60 million reads with RNase R enrichment. For comprehensive low-abundance circRNA discovery, up to 150 million reads per sample.

Can circRNA-seq detect circRNAs from FFPE samples?

Yes, provided DV200 ≥ 40%. RNase R treatment is not recommended for FFPE RNA. Deep sequencing (≥100 M reads) is recommended for circRNA detection from FFPE samples.

Circular RNA Sequencing (circRNA-seq) Service | Back-Splice Junction Detection and circRNA Profiling

Circular RNA Sequencing Overview

Circular RNAs are generated through a non-canonical splicing event called back-splicing, in which a downstream 5' splice site is joined to an upstream 3' splice site, creating a covalently closed loop structure. This unique architecture renders circRNAs resistant to RNase R digestion and other exonucleases, contributing to their exceptional stability compared with linear transcripts. Since the initial discovery that circRNAs are pervasively expressed across human cell types (Salzman et al. 2012), thousands of circRNAs have been annotated, and their regulatory functions in sequestering microRNAs, interacting with RNA-binding proteins, and modulating gene expression have been increasingly recognised.

Detecting and quantifying circRNAs presents unique challenges not shared with linear RNA analysis. Standard RNA-seq library preparation using poly(A) capture fails to enrich circRNAs because they lack polyA tails. Moreover, conventional RNA-seq alignment pipelines discard reads that map discontinuously across the genome — precisely the reads that characterise back-splice junctions. Our circRNA-seq service addresses these challenges through rRNA depletion (not polyA capture), RNase R treatment for optional linear RNA removal, and dedicated bioinformatic algorithms specifically designed for back-splice junction detection. For comprehensive analysis of both circular and linear transcripts from the same library, we recommend our Total RNA-Seq service, which can be analysed for both linear and circular RNA expression. For exosome-derived circRNAs, our Exosomal circRNA Sequencing service provides specialised protocols optimised for biofluid samples.

Feature	circRNA-seq (Our Service)	Total RNA-Seq	mRNA-Seq (PolyA)
RNA Enrichment Method	rRNA depletion ± RNase R treatment	rRNA depletion only	Poly(A) capture
circRNA Detection	Yes — optimised with optional RNase R enrichment	Yes — standard detection	No — circRNAs lack polyA tails
Linear RNA Detection	Limited (with RNase R) / Yes (without RNase R)	Yes — all linear RNAs preserved	Yes — polyadenylated only
circRNA Detection Sensitivity	High — RNase R removes competing linear RNA background	Moderate — linear RNA may mask low-abundance circRNAs	None — circRNAs not captured
Back-Splice Junction Read Ratio	High — linear RNA depletion enriches junction reads	Low — only 0.1–1% of reads span back-splice junctions	N/A
circRNA Isoform Resolution	Good — dedicated circRNA aligners identify exact junction coordinates	Moderate — depends on sequencing depth	N/A
mRNA / lncRNA Quantification	No (with RNase R) / Yes (without)	Yes — comprehensive	Yes — focused on polyA transcripts
Recommended Read Length	PE150	PE150	PE150
Best For	circRNA-focused studies; novel circRNA discovery; low-abundance circRNA detection; circRNA isoform analysis; ceRNA network construction	Comprehensive coding + non-coding transcriptome analysis including circRNAs as part of a broader survey	Protein-coding gene expression profiling; studies where non-coding RNAs are not of primary interest

Technical Advantages of Our circRNA-seq Service

RNase R Enrichment for Enhanced circRNA Detection

Circular RNAs are resistant to RNase R, a 3' to 5' exoribonuclease that degrades linear RNA molecules. Our optional RNase R pre-treatment step selectively digests linear RNA (mRNA, lncRNA, rRNA fragments) before library preparation, enriching the circRNA fraction by 10–50 fold. This dramatically increases the proportion of back-splice junction reads in the sequencing library, enabling the detection of low-abundance circRNAs that would otherwise be missed. For studies where both circular and linear RNA quantification are required, we offer the option to split each sample into paired RNase R-treated and untreated libraries, enabling parallel circRNA and linear transcript profiling from the same biological material.

Multi-Algorithm circRNA Detection Pipeline

Accurate circRNA identification requires specialised bioinformatic tools capable of detecting back-splice junction reads. Rather than relying on a single algorithm, we apply a consensus-based approach using multiple independent circRNA detection tools (CIRI2, CIRCexplorer2, and find_circ), each employing different mapping strategies and filtering criteria. Only circRNA candidates identified by at least two algorithms are reported, significantly reducing false-positive rates. Each candidate is further filtered based on read support (minimum 2 unique back-splice junction reads), junction quality scores, and alignment consistency. Our pipeline also distinguishes between exonic circRNAs, intronic circRNAs (EIciRNAs), and intergenic circRNAs, providing comprehensive annotation of the circular transcriptome.

Integrated Functional and Network Analysis

Beyond circRNA identification and quantification, our bioinformatics pipeline provides deep functional interpretation. We construct ceRNA (competing endogenous RNA) networks by predicting miRNA response elements (MREs) on identified circRNAs and linking them to their target mRNAs. Host gene functional enrichment analysis reveals the biological pathways associated with differentially expressed circRNAs. Where matched RNA-seq data are available, we perform integrated analysis to examine how circRNA expression changes correlate with host gene expression and broader transcriptomic alterations. All results are delivered with publication-ready visualisations including circRNA genomic maps, back-splice junction read validation plots, circRNA-miRNA-mRNA network diagrams, and expression heatmaps.

Bioinformatics and Data Analysis

Our circRNA bioinformatics pipeline is designed for sensitive and specific detection of back-splice junctions, accurate quantification of circRNA expression levels, and comprehensive functional interpretation of the circular transcriptome.

Analysis Package	Content Description
Standard Analysis
1. Raw Data QC and Preprocessing	FastQC quality assessment, adapter trimming (Cutadapt/Fastp), rRNA content assessment, read filtering (Q ≥ 30). Comprehensive QC report including per-base quality, GC content, duplication rates, and back-splice junction discovery rates.
2. Read Alignment and Chimeric Junction Detection	Alignment to reference genome using STAR with chimeric junction detection enabled. Identification of discordant read pairs where one mate maps in the forward orientation and the other in the reverse orientation, spanning the back-splice junction. Linear alignment statistics and genome-wide coverage tracks.
3. circRNA Identification and Quantification	Multi-algorithm circRNA detection using CIRI2 (CIGAR-based junction identification), CIRCexplorer2 (annotation-guided back-splice mapping), and find_circ (anchor-based detection). Consensus filtering: only circRNAs supported by ≥2 algorithms and ≥2 unique back-splice junction reads are reported. Quantification in RPM (reads per million mapped reads) and normalised counts for differential analysis.
4. circRNA Annotation and Classification	Annotation of identified circRNAs by genomic origin: exonic circRNAs (back-splice within known gene boundaries), circular intronic RNAs (EIciRNAs, intron-retained circRNAs), intergenic circRNAs (originating from intergenic regions), and antisense circRNAs. Assignment to host genes and transcript biotypes. Comparison with public circRNA databases (circBase, CircAtlas, CSCD).
5. Differential circRNA Expression Analysis	Identification of differentially expressed circRNAs between experimental groups using DESeq2 (on back-splice junction counts) or edgeR. Normalisation for library size and sequencing depth. Multiple testing correction (FDR < 0.05). Separate analysis of circRNA-type-specific differential expression. Visualisation: volcano plots, MA plots, heatmaps.
Advanced Analysis
6. circRNA-MicroRNA Interaction and ceRNA Network Analysis	Prediction of miRNA response elements (MREs) on identified circRNA sequences using miRanda, TargetScan, and RNAhybrid. Construction of circRNA-miRNA-mRNA ceRNA networks. Identification of circRNAs with high miRNA sponge potential (multiple MREs for disease-relevant miRNAs). Network visualisation and hub circRNA identification.
7. circRNA Isoform and Alternative Back-Splicing Analysis	Identification of multiple circRNA isoforms originating from the same host gene locus (alternative back-splicing events). Circular-to-linear (C/L) ratio calculation for each circRNA-host gene pair. Detection of circular isoform switching events between conditions. Characterisation of circRNA-specific exon usage patterns.
8. circRNA-RBP Interaction Analysis	Prediction of RNA-binding protein (RBP) binding sites on identified circRNA sequences using curated RBP motif databases (RBPsuite, ATtRACT). Identification of circRNA-RBP interaction networks. Enrichment analysis of RBP types among differentially expressed circRNAs. Integration with CLIP-seq data where available.
9. Integrative and Multi-Omics Analysis	Integration of circRNA expression data with matched mRNA-seq, miRNA-seq, or proteomics data. Correlation analysis between circRNA expression and host gene expression. Functional enrichment of host genes of differentially expressed circRNAs (GO, KEGG, Reactome). Survival analysis and clinical correlation (where clinical metadata are available).

Our analysis team delivers a comprehensive report with publication-ready figures including genome-wide circRNA distribution maps, back-splice junction validation plots (IGV sashimi plots), circRNA expression heatmaps, ceRNA network diagrams, and functional enrichment visualisations. Data are provided in standard formats (circRNA coordinate BED files, junction read count matrices, RPM expression tables, differential expression results, and network files compatible with Cytoscape).

Applications

Circular RNA sequencing has broad applications across basic biology, translational research, and biomarker development. The following areas are particularly well-suited to our circRNA-seq approach.

circRNA Biomarker Discovery in Liquid Biopsy and Tissue

The exceptional stability of circRNAs — conferred by their closed-loop structure that resists exonuclease degradation — makes them ideal biomarker candidates in biofluids and tissue samples. CircRNAs have been identified in plasma, serum, urine, and cerebrospinal fluid, often at higher stability than linear RNA counterparts. Total RNA-seq of biofluid-derived samples can identify differentially expressed circRNAs that distinguish disease states, including cancer, cardiovascular disease, and neurodegenerative disorders. Our circRNA-seq service, combined with bioinformatic ceRNA network analysis, provides a comprehensive platform for circRNA biomarker discovery and validation.

Functional Characterisation of circRNA Mechanisms

CircRNAs exert their biological functions through multiple molecular mechanisms: (1) miRNA sponging — sequestering miRNAs away from their mRNA targets through shared miRNA response elements, as exemplified by the well-characterised ciRS-7/CDR1as circular RNA that sponges miR-7; (2) RNA-binding protein (RBP) interaction — serving as scaffolds or decoys for RBPs to modulate their activity; (3) transcriptional regulation — interacting with RNA polymerase II or other transcriptional machinery at their host gene loci; (4) translation templates — a subset of circRNAs containing internal ribosome entry sites (IRES) or m6A modifications can be translated into functional micropeptides. Our integrated bioinformatic analysis addresses all these functional dimensions.

ceRNA Regulatory Network Construction

CircRNAs often function as competing endogenous RNAs (ceRNAs) that sequester miRNAs, thereby regulating the expression of miRNA target mRNAs. Our circRNA-seq analysis pipeline includes comprehensive ceRNA network construction: miRNA response element (MRE) prediction on identified circRNA sequences using multiple algorithms (miRanda, RNAhybrid, TargetScan), integration with mRNA expression data from matched samples, statistical modelling of ceRNA interaction likelihood, and network topology analysis. This approach has been successfully applied to identify circRNA-mediated regulatory hubs in cancer, cardiac disease, and developmental biology, and is particularly powerful when combined with parallel miRNA-seq or small RNA-seq data.

Cancer Transcriptomics and Tumour Biology

CircRNA dysregulation is increasingly recognised as a hallmark of cancer, with specific circRNAs functioning as oncogenes or tumour suppressors. Large-scale circRNA profiling studies have identified thousands of cancer-specific circRNAs, with many showing diagnostic or prognostic value. Our circRNA-seq service is applicable to a wide range of cancer research contexts: (1) comprehensive circRNA profiling of tumour vs. normal tissue to identify cancer-specific circRNAs; (2) correlation of circRNA expression with clinical parameters (stage, grade, survival); (3) identification of circRNA expression signatures associated with treatment response or resistance; (4) integration with genomic data to identify circRNA-generating fusion genes. Recent studies have highlighted the global suppression of circRNA biogenesis by oncogenic transcription factors such as MYCN, revealing new layers of circRNA regulation in cancer (Fuchs et al. 2023).

Developmental Biology, Neuroscience, and Disease Mechanism Studies

CircRNAs are highly enriched in the brain compared to other tissues, with many showing dynamic expression patterns during neuronal differentiation and development. Neural circRNAs are often derived from synaptic genes and are thought to regulate local translation at synapses. In developmental biology, circRNAs exhibit tissue-specific and developmental stage-specific expression, suggesting important roles in cell fate determination and organogenesis. Our circRNA-seq service enables researchers to profile the circular transcriptome across developmental time courses, identify neural-specific circRNAs, and investigate circRNA function in model systems. For neuroscience studies, our long-read sequencing partners also offer nanopore circRNA sequencing for full-length circRNA isoform characterisation, complementary to our short-read circRNA-seq service for quantitative profiling.

Sample Requirements

Sample Type	Recommended Amount	Quality Requirements
Total RNA (standard)	1–5 µg (with RNase R); 100 ng – 1 µg (without)	RIN ≥ 7; A260/A280 ≥ 1.8; A260/A230 ≥ 1.5; DNA-free
Total RNA (FFPE / degraded)	1–5 µg	DV200 ≥ 40%; RNase R treatment not recommended for highly degraded RNA
Fresh-frozen tissue	20–50 mg	Snap-frozen in liquid nitrogen; stored at −80°C
FFPE tissue sections	5–10 sections (10 µm thick)	RNA extraction performed in-house; RNase R treatment assessed case-by-case
Biofluid (plasma/serum)	1–3 mL	EDTA or citrate tubes; haemolysis-free; RNase R treatment not needed (low linear RNA background)
Cultured cells	5 × 10⁶ – 1 × 10⁷ cells	Washed in PBS; snap-frozen or TRIzol-lysed; RNA stabilisation reagent accepted

Important Notes:

RNase R treatment requires higher input RNA (≥1 µg) due to partial RNA loss during enzymatic digestion. For limited samples, we recommend proceeding without RNase R and relying on deep sequencing depth for circRNA detection.
For studies comparing circRNA expression across multiple conditions, we strongly recommend using consistent RNase R treatment conditions across all samples to avoid batch effects.
Minimum 3 biological replicates per experimental group is recommended for robust statistical analysis of circRNA differential expression, given the typically low expression levels of individual circRNAs.
For circRNA biomarker discovery in biofluids, we recommend including matched RNase R-treated and untreated samples from at least one pilot sample to determine the optimal strategy.
Genomic DNA contamination must be removed (DNase I treatment) as gDNA can generate false-positive back-splice junction calls from intergenic or trans-splicing artefacts.
Samples should be shipped on dry ice or in RNA stabilisation reagent. Avoid multiple freeze-thaw cycles.

Case Study: Whole-Transcriptome circRNA Profiling Reveals MYCN-Dependent Suppression of Circular RNA Biogenesis in Neuroblastoma

A 2023 study published in Nature Communications by Fuchs and colleagues performed deep whole-transcriptome sequencing to characterise the circRNA landscape in 104 primary neuroblastomas across all clinical risk groups, uncovering a global suppressive function of the MYCN oncogene on circRNA biogenesis.

Neuroblastoma is the most common extracranial solid tumour in children, with highly variable clinical outcomes ranging from spontaneous regression to aggressive, treatment-resistant disease. MYCN amplification is the strongest genetic marker of poor prognosis in neuroblastoma, driving a transcriptional program that promotes proliferation, dedifferentiation, and metastasis. While MYCN's effects on the linear transcriptome were well characterised, its impact on the circular RNA landscape — and the broader relationship between circRNA expression and neuroblastoma biology — remained largely unknown. Given the emerging roles of circRNAs in cancer, understanding how oncogenic drivers such as MYCN shape the circular transcriptome could reveal new biomarkers and therapeutic targets.

Figure 1. Study design for whole-transcriptome circRNA sequencing of neuroblastoma.
Total RNA was extracted from 104 primary neuroblastoma tissues spanning all clinical risk groups (high-risk MYCN-amplified, high-risk non-amplified, intermediate-risk, low-risk). Ribosomal RNA was depleted, and strand-specific libraries were sequenced on Illumina NovaSeq (PE150). Bioinformatic analysis identified circRNAs using CIRCexplorer2 and CIRI2, with validation by RNase R treatment, qRT-PCR, and Sanger sequencing. Functional experiments confirmed the MYCN-DHX9-circRNA regulatory axis. Adapted from Fuchs et al. 2023 (CC BY 4.0).

Whole-transcriptome circRNA sequencing approach: The authors collected 104 primary neuroblastoma tumour specimens representing all clinical risk groups, including MYCN-amplified (MNA) and non-amplified tumours. Total RNA was extracted from snap-frozen tissues and subjected to rRNA depletion (Ribo-Zero Gold) to remove cytoplasmic and mitochondrial rRNAs. Strand-specific RNA-seq libraries were constructed using a dUTP-based approach and sequenced on Illumina NovaSeq 6000 with PE150 read length, generating 60–100 million reads per sample. circRNA identification was performed using two independent algorithms: CIRCexplorer2 (which uses an annotation-guided approach to identify back-splice junctions) and CIRI2 (which detects junction-spanning reads from CIGAR alignment strings). Only circRNAs detected by both tools were retained for downstream analysis. Validation experiments included RNase R treatment (to confirm circular structure), qRT-PCR with divergent primers, and Sanger sequencing of back-splice junction products. Functional experiments involved MYCN knockdown and DHX9 perturbation in neuroblastoma cell lines (IMR-32, NGP, SK-N-SH) followed by circRNA expression analysis. The study also integrated RNA-seq, ChIP-seq (H3K27ac, MYCN), and ATAC-seq data to characterise the chromatin landscape around circRNA-producing genes.

MYCN suppresses circRNA biogenesis through DHX9-mediated regulation of back-splicing Figure 2. MYCN globally suppresses circRNA biogenesis through DHX9.
Neuroblastomas with MYCN amplification showed globally reduced circRNA expression compared with non-amplified tumours. MYCN directly upregulates DHX9, an RNA helicase that binds to inverted Alu repeats flanking circRNA-producing exons and inhibits back-splicing. Knockdown of MYCN or DHX9 rescued circRNA expression. The circARID1A circRNA was identified as significantly upregulated in neuroblastoma and promotes tumour cell growth. Adapted from Fuchs et al. 2023 (CC BY 4.0).

Key findings: (1) circRNA expression was systematically lower in MYCN-amplified (MNA) neuroblastomas compared with non-amplified tumours, revealing a global suppressive effect of MYCN on circRNA biogenesis. (2) DHX9, an RNA helicase that resolves RNA secondary structures, was identified as a key mediator of MYCN-dependent circRNA suppression. MYCN directly binds the DHX9 promoter and activates its transcription. DHX9 then binds to inverted Alu repeat elements flanking circRNA-producing exons, destabilising base-pairing interactions required for back-splicing and thereby inhibiting circRNA formation. (3) This MYCN-DHX9-circRNA axis was not unique to neuroblastoma — the same suppressive effect was observed in MYCN-amplified medulloblastoma, suggesting a general mechanism by which oncogenic transcription factors suppress circRNA biogenesis. (4) Despite the global suppression, 25 circRNAs were specifically upregulated in neuroblastoma relative to other cancer types. Among these, circARID1A (derived from the ARID1A tumour suppressor gene) was identified as a functionally important circRNA that promotes neuroblastoma cell growth and survival through direct interaction with the KHSRP RNA-binding protein. (5) circARID1A expression was associated with poor prognosis in neuroblastoma, and its growth-promoting function was confirmed by knockdown experiments in multiple cell lines. This study demonstrates the power of whole-transcriptome circRNA sequencing for discovering clinically relevant circRNA regulatory mechanisms and identifying functionally important circular transcripts in cancer.

FAQs — Frequently Asked Questions

- 1) What is a circular RNA and how is it different from linear RNA?
  - A circular RNA (circRNA) is a covalently closed RNA molecule formed by back-splicing, where a downstream 5' splice site joins to an upstream 3' splice site, creating a continuous loop structure. Unlike linear RNAs, circRNAs lack 5' and 3' ends (and therefore lack caps and polyA tails), making them resistant to exonuclease degradation. This closed structure also means circRNAs are not affected by RNase R treatment, which degrades linear RNAs. Most circRNAs are generated from exons of protein-coding genes, though some originate from introns (EIciRNAs) or intergenic regions. Their circular architecture enables unique functions including miRNA sponging, RBP scaffolding, and, in some cases, cap-independent translation.
- 2) Why is RNase R treatment important for circRNA-seq?
  - RNase R is a 3' to 5' exoribonuclease that selectively degrades linear RNA molecules while leaving circular RNAs intact due to their closed-loop structure. Treating total RNA with RNase R before library preparation enriches the circRNA fraction by removing the abundant linear RNA background (mRNA, lncRNA, rRNA). This dramatically increases the proportion of back-splice junction reads in the sequencing data (from ~0.1–1% to >10–50% of total reads), enhancing the detection sensitivity for low-abundance circRNAs. However, RNase R treatment is optional and has trade-offs: it requires higher RNA input (due to RNA loss during digestion), can partially digest some circRNAs if over-treated, and prevents simultaneous quantification of linear transcripts from the same library. We recommend RNase R treatment for circRNA-focused discovery studies but not for studies requiring parallel linear transcript analysis.
- 3) How do you distinguish genuine circRNAs from false positives?
  - We employ multiple layers of false-positive filtering: (1) Multi-algorithm consensus — only circRNAs detected by ≥2 independent detection tools (CIRI2, CIRCexplorer2, find_circ) are reported, as each tool uses different mapping strategies with distinct false-positive profiles. (2) Read support threshold — minimum 2 unique back-splice junction reads, with additional filtering for junction read quality and mapping scores. (3) Alignment artefact filtering — reads from template-switching during reverse transcription, trans-splicing artefacts, and genomic DNA rearrangements are computationally identified and removed. (4) Cross-sample reproducibility — circRNAs detected in only one sample without reproducible expression patterns are flagged as low-confidence. (5) Experimental validation — top candidates can be validated by RNase R resistance confirmation, divergent primer RT-PCR, and Sanger sequencing of back-splice junctions.
- 4) What sequencing depth is recommended for circRNA-seq?
  - For standard circRNA detection and differential expression analysis without RNase R treatment, we recommend 80–100 million reads per sample (PE150). With RNase R enrichment, 40–60 million reads are typically sufficient due to the higher proportion of back-splice junction reads. For comprehensive discovery of low-abundance circRNAs or full characterisation of the circular transcriptome, we recommend up to 150 million reads per sample. The required depth also depends on the tissue type (brain samples typically have higher circRNA diversity and require less depth), sample quality, and the specific research question. We provide sequencing depth recommendations during project planning based on your sample types and research objectives.
- 5) Can circRNA-seq detect circRNAs from low-input or FFPE samples?
  - Yes, but with important caveats. For FFPE samples, circRNA detection is feasible provided the RNA quality meets DV200 ≥ 40%. However, RNase R treatment is not recommended for FFPE RNA as the already-fragmented RNA may result in excessive loss of material and reduced library complexity. For low-input samples, we recommend proceeding without RNase R treatment and instead relying on deep sequencing (≥100 M reads) for circRNA detection. The strand-specific library preparation protocol is compatible with 10–100 ng of total RNA input. For biofluid samples (plasma/serum), RNase R treatment is generally not needed as the circulating RNA fraction intrinsically contains a higher proportion of circular RNA due to the selective degradation of linear RNA in the extracellular environment.
- 6) How do you quantify circRNA expression levels?
  - CircRNA expression is quantified using back-splice junction reads — these are sequencing reads that span the unique junction where the 3' end joins to the 5' end of the circular RNA. Expression is reported as RPM (reads per million mapped reads) or as raw junction read counts. Importantly, because circRNAs do not have a defined 5' end, standard gene expression metrics like FPKM/RPKM (which normalise by transcript length) are not directly applicable. We also calculate the circular-to-linear (C/L) ratio, which compares back-splice junction read density to canonical (linear) splice junction read density for the same gene, providing a normalised measure of circRNA production efficiency. For differential expression analysis between groups, we use DESeq2 on back-splice junction read counts, with appropriate normalisation for sequencing depth and library composition.
- 7) Can circRNA-seq data be used to predict circRNA functions?
  - Yes — our bioinformatics pipeline provides multiple layers of functional prediction: (1) miRNA sponge prediction — we identify miRNA response elements (MREs) on identified circRNAs and construct ceRNA networks linking circRNAs to their potential miRNA targets and downstream mRNAs. (2) RBP interaction prediction — we scan circRNA sequences for RNA-binding protein motif matches and predict circRNA-RBP interactions that may mediate scaffolding or decoy functions. (3) Host gene correlation — we analyse the relationship between circRNA expression and host gene linear expression, which can indicate co-regulation or competitive biogenesis. (4) Coding potential assessment — for circRNAs containing internal ribosome entry sites (IRES) or predicted m6A modifications, we evaluate protein-coding potential using tools such as CPC2 and CNCI. These functional predictions provide a foundation for designing targeted validation experiments.
- 8) What is the relationship between circRNA-seq and other RNA interaction methods?
  - circRNA-seq is complementary to several other RNA analysis approaches available through our service portfolio. Total RNA-Seq provides broader transcriptome coverage including both linear and circular RNAs from the same library, while circRNA-seq with RNase R treatment offers deeper circRNA-specific coverage. Small RNA Sequencing profiles miRNAs, which can be integrated with circRNA-seq data for ceRNA network construction. lncRNA Sequencing focuses on linear lncRNAs, many of which share host genes with circRNAs — integrated analysis can reveal competitive splicing regulation. For functional studies following circRNA discovery, we recommend combining circRNA-seq with RNA-seq after circRNA perturbation to identify downstream target genes and pathways.

References:

Salzman J, Gawad C, Wang PL, et al. Circular RNAs are the predominant transcript isoform from hundreds of human genes in diverse cell types. PLoS ONE. 2012;7(2):e30733.
Zhang J, Chen S, Yang J, Zhao F. Accurate quantification of circular RNAs identifies extensive circular isoform switching events. Nature Communications. 2020;11:90.
Fuchs S, Danßmann C, Klironomos F, et al. Defining the landscape of circular RNAs in neuroblastoma unveils a global suppressive function of MYCN. Nature Communications. 2023;14:3936.
Rahimi K, Venø MT, Dupont DM, Kjems J. Nanopore sequencing of brain-derived full-length circRNAs reveals circRNA-specific exon usage, intron retention and microexons. Nature Communications. 2021;12:4825.

For Research Use Only. This service is intended for circular RNA discovery, characterisation, and functional analysis applications. It is not intended for clinical diagnosis, treatment selection, patient stratification, or therapeutic decision-making.

Circular RNA Sequencing (circRNA-seq) Service

Circular RNA Sequencing Overview

circRNA-seq vs. Total RNA-Seq vs. mRNA-Seq: circRNA Detection Capability