Talanta 238 (2022) 123066

Contents lists available at ScienceDirect

Talanta
journal homepage: www.elsevier.com/locate/talanta

Circular RNA detection methods: A minireview

Zhang Mi a, Chen Zhongqiang b, Jiang Caiyun c, Liu Yanan a, Wu Jianhua a, Liu Liang a, *
a
Department of Pharmacy, Zhongnan Hospital of Wuhan University, Wuhan, 430071, China
b
School of Medicine, Jianghan University, Wuhan, 430056, China
c
Department of Pharmacy, The Third Affiliate Hospital of Sun Yat-Sen University, Guangzhou, 510630, China

A R T I C L E I N F O A B S T R A C T

Keywords: Circular RNA (circRNA), a novel type of covalently closed RNA, is implicated in several developmental and
circRNA detection metabolic disease processes. CircRNAs exhibit tissue-specific expression, and are stable, abundant, and highly
Novel method conserved, making them ideal biomarkers for diagnosis and prognosis. Accurate profiling of circRNA, however, is
Biomarker
a prerequisite for their clinical application. Traditional methods such as northern blotting, RT-qPCR, and
Amplification
Advantages and limitations
microarray analysis provide useful but limited information. To address these issues, a number of novel assays
have recently emerged, such as droplet digital PCR (ddPCR), isothermal exponential amplification, and rolling
cycle amplification, which increase the sensitivity and specificity of circRNA detection. Herein, we summarize
the advantages and limitations of the new detection methods and discuss the challenges as well as future
directions.

1. Introduction methods for detecting circRNAs, such as northern blotting, qRT-PCR,

microarray analysis, and RNA sequencing, fall short of the demands of
Circular RNA (circRNA) is a novel class of non-coding RNA that modern research. Northern blotting can be used to evaluate the presence
consists of at least a few hundred nucleotides covalently bound in a and size of circRNA but has low sensitivity and is labor-intensive and
closed loop that lacks 5ʹ -3ʹ polarity and a poly-A tail [1,2]. Generally, time-consuming [19]. qRT-PCR is the only assay available that can
circRNAs are generated by cyclization of specific RNAs via back splicing readily quantify circRNA. Reverse transcription of circRNA to generate
[3,4]. Accumulated evidence has shown that circRNAs can act as cDNA for qRT-PCR, however, can result in template-switching artifacts,
sponges for microRNAs (miRNAs), offer platforms for assembly of RNA leading to an overestimation of circRNA concentration [20–22]. RNA
binding proteins, and interact with mRNAs to regulate their expression sequencing, another widely used method in circRNA research, is labor
post-transcriptionally [5–9]. Recent studies have shown that circRNAs intensive and requires costly reagents and equipment, as well as trained
exhibit high expression, tissue specificity, structural stability, and high personnel for data processing, limiting its usefulness in routine diagnose
conservation [10–12]. Based on these characteristics, circRNAs have the [23,24]. Microarray-based approaches generate data that are difficult to
potential to serve as vital biomarkers for diagnosing and predicting compare between studies; thus, those techniques are more suitable as
disease progression and prognosis [1,13–16]. For use in clinical appli­ screening tools rather than for quantification [25]. Therefore, there is an
cations, it is necessary to evaluate which circRNAs are altered in a urgent demand to develop simple, sensitive, stringent and effective
specific disease to ensure whether these alterations are representative of methods for circRNA analysis.
that disease. Importantly, accurate detection of circRNA is a prerequisite Recent research has focused on developing more efficient and low-
for its clinical applications. cost approaches with superior flexibility and adaptability for circRNA
Owing to its circular structure, circRNA is difficult to separate from analysis. Approaches currently being developed include reverse
other RNA species through commonly used techniques, such as elec­ transcription-droplet digital polymerase chain reaction (RT-ddPCR),
trophoresis or through methods targeting poly-A tails [17]. Moreover, rolling cycle amplification (RCA), and stem-loop primer-induced double
endogenous circRNAs have relatively low abundance as compared to exponential amplification (LAMP), as shown in Fig. 1. Herein, we
their linear counterparts [18]. These characteristics pose a challenge for describe new techniques available for circRNA detection and discuss
the analysis of circRNAs in the research community. Traditional their advantages and limitations for improving circRNA detection.

* Corresponding author. Department of Pharmacy, Zhongnan Hospital of Wuhan University, 169 Donghu Road, Wuchang District, Wuhan, 430071, China.
E-mail address: [email protected] (L. Liang).

https://doi.org/10.1016/j.talanta.2021.123066
Received 16 September 2021; Received in revised form 11 November 2021; Accepted 12 November 2021
Available online 16 November 2021
0039-9140/© 2021 Elsevier B.V. All rights reserved.
Z. Mi et al. Talanta 238 (2022) 123066

system (32P) [33]. Other probe-design strategies, such as using lock

nucleic acid (LNA) oligonucleotide probes to substitute traditional DNA
oligonucleotide probes, have been shown to have at least a 10-fold
higher efficiency than traditional DNA probes for small RNA [34,35].
The LNA oligonucleotide is a new class of bicyclic RNA analog that
demonstrates an unprecedented high affinity for its complementary
DNA or RNA targets. This strategy may be also suitable for improving
the detection sensitivity in circRNA northern blot detection.

2.1.2. Microarray-based circRNA detection

With the development of new techniques, increasing numbers of
circRNAs are being discovered. A systematic investigation of the profiles
of these circRNAs in disease is required before they can be used as
biomarkers in clinical practice. Such an investigation will require a high-
throughput method that can detect large amounts of circRNAs.
Microarray-based methods are particularly effective high-throughput
tools that can simultaneously and comprehensively detect a wide
range of circRNAs [25,36]. Microarray assays for circRNA are similar to
those for linear mRNAs, differing only in the design of the probe se­
quences [37], which targets the ‘back-splice’ junction site sequence
within target circRNAs [38].
To perform microarray analysis for circRNA, total RNA is isolated
from tissues or cells and then digested by RNase R to remove linear RNA.
After digestion, amplification and transcription are performed using a
random priming method. Labeled cRNAs are then hybridized onto a
Fig. 1. Schematic representation of different circRNA profiling methods.
commercially available circRNA array. The hybridized microarray slides
are then washed and scanned using a fluorescence scanner. The types
2. Methods of circRNA detection and relative quantities of circRNAs can be obtained from the fluores­
cence signal intensity.
The association between abnormal expression of circRNAs and the Commercially available circRNA microarrays include the Arraystar,
initiation and development of diseases is well supported [26–28]. CapitalBio Technology, and CD microarrays. The major differences in
Although circRNA has features that make it a suitable biomarker in the different platforms involve the design of the probe, immobilization
clinical applications, its analysis is challenging because of its low con­ strategy, sample labeling, and signal-detection system. To acquire ac­
centration, circular structure, and lack of 5ʹ-3ʹ polarity and a polyA tail curate and reliable results, each platform uses a group of exogenous RNA
[29]. Considering the pivotal role that circRNAs play in different ap­ controls across the samples.
plications, significant progress has been made over the past decade to­ CircRNA microarrays are powerful tools for the identification of
ward building new profiling platforms. differentially expressed circRNAs in large-scale parallel assessments.
The detection efficiency of circRNA microarrays is superior to that of
2.1. Hybridization-based methods RNA-sequencing, as almost 100% of detected signals derived from the
microarray could be applied for profiling circRNAs [39]. This technique,
2.1.1. Northern blotting however, requires a high input of total RNA, can only measure relative
Validation of the presence of specific circRNAs is of great importance concentrations of RNA, and produces results that are difficult to
when studying circRNA. RT-PCR assays that use RNase R, a 3ʹ-5ʹ compare between studies. Therefore, microarray analysis is commonly
exonuclease, to digest and remove linear RNAs provide the initial used for screening but not quantifying circRNAs. Theoretically, non­
experimental evidence for RNA circularity. RNase R, however, can labeling and non-amplification reaction detection methods are simpler
digest large molecular weight circRNA and may fail to digest linear and the acquired results are more credible, especially for diagnostic
RNAs with stable structures [30]. Northern blotting reveals the size and goals. Thus, future works should mainly focus on two aspects: first, la­
circular configuration of RNAs [31] and is currently the gold standard beling the secondary probe rather than the circRNA target to increase
for confirming the presence of circRNAs, but it is limited by low sensi­ the sensitivity, and second, making use of special equipment to decrease
tivity, low throughput, and time-consuming steps [32]. the input of total RNA.
A typical northern blotting protocol uses gel electrophoresis to
separate denatured RNAs. RNAs are then transferred to a nitrocellulose 2.1.3. Deep sequencing
membrane by capillary or electrical methods. After transfer, RNA is Owing to their potential significant roles in disease diagnosis and
covalently linked to the membrane through baking or exposure to ul­ prognosis, researchers are showing increasing interest in the genome-
traviolet light. Hybridization is then performed between a fluorescent or wide expression patterns of circRNAs. Advances in sequencing tech­
radiolabeled probe and the RNAs are immobilized on the membrane. nology, particularly the development of deeper sequencing with longer
Hybrid signals are profiled according to the method of probe labeling. read lengths, have accelerated genome-wide studies of circRNA [40].
Notably, circRNAs have a lower apparent mobility than straight RNA in Furthermore, algorithms for mapping RNA to its genomic source and the
denaturing polyacrylamide gels, while having the same running strategies for ribosomal RNA (rRNA) depletion have enabled the
behavior as similarly sized straight RNAs in agarose gels. sequencing of non-polyadenylated RNA [41]. The various approaches
Numerous teams have attempted to improve the sensitivity and for library enrichment in circRNAs generally fall into two broad stra­
reduce the total assay time of northern blots for circRNAs. For example, tegies [42,43]. The first option, known as candidate-based approach, is
Xiaolin Wang and Ge Shan have reported the use of digoxigenin (DIG)- to generate a list of candidate junctions using the existing transcript
labeling system for the detection of circRNA. This detection system has models [6,10]. This approach has been used to acquire reliable results;
several advantages: high sensitivity, short exposure time, longer shelf moreover, it is the fastest for implementation in ribosomal
life, and increased safety compared with the traditional isotope labeling RNA-depleted libraries but cannot be used to detect novel circRNAs

2
Z. Mi et al. Talanta 238 (2022) 123066

present in the samples. The second option identifies junctions by laboratories.

matching RNA sequencing reads to the genomic sequence, as is done by
spliced alignment algorithms [44]. In this approach, libraries are 2.2. Amplification-based methods
rRNA-depleted and treated with RNase R treatment before
high-throughput sequencing. Sequence reads are then mapped directly 2.2.1. PCR-based circRNA detection
to de novo genomic positions, allowing the identification of back spliced RT-PCR was first applied in the validation of circRNA obtained from
reads in specific sequences. This second approach avoids the bias RNA sequencing, and it remains to be the most used assay for circRNA
introduced by using the existing models to generate a candidate list and profiling. qRT-PCR, the “gold standard” for quantification of circRNA,
permits the identification of novel circRNAs via the detection of unan­ covers a huge dynamic range and is approximately one thousand-fold
notated splice junctions. However, it is less accurate, less sensitive, and more sensitive than hybridization-based approaches [55,56]. In a
requires a higher amount of total RNA than the candidate-based common qRT-PCR protocol, total RNA is isolated from cells or tissues
approach. In addition, it is more prone to be influenced by endonu­ and then treated with RNase R to remove linear RNAs. This is followed
clease contamination. by RT to obtain complementary DNA (cDNA) for circRNA. The cDNA is
With the abundance of RNA sequence data, numerous algorithms for PCR amplified using a primer to a specific splice junction. Finally, the
calculating and visualizing circRNA have been recently developed. For fluorescence signals are determined and analyzed. This technology
example, Li Chen built the PcircRNA_finder software based on the represents a balance of sensitivity, specificity, speed, cost, and
analysis of simulated and real rRNA-/RNAase R RNA-Seq data of Ara­ simplicity, which makes qRT-PCR-based techniques easy to carry out in
bidopsis and rice, which shows a good comprehensive ability and greater a clinical laboratory [56]. Those advantages offer the possibility of early
sensitivity and precision in predicting circRNAs in plants [45]. PcircR­ diagnosis and monitoring of treatment efficacy. While qRT-PCR tech­
NA_finder is mainly designed for exonic circRNA prediction and consists nology has achieved sensitive determination of circRNA, exonuclease
of three modules: catcher, annotator and filter. Niu et al. developed a RNase R may have insufficient digestion of linear RNA with a secondary
classification system called CirRNAPL, which extracts the features of structure at the 3ʹ end [57]. Moreover, a single circRNA molecule can
nucleic acid composition and structure of the circRNA sequence and produce several repeated cDNA copies due to strand displacement and
optimizes the extreme learning machine based on the particle swarm rolling circle replication during reverse transcription, and these artifacts
optimization algorithm, providing significantly improved identification can account for 34–55% of the RNA detected, leading to an over­
accuracy compared with blast [46]. Other computational methods are estimation of circRNA concentration [20].
available such as pipeline CIRCexplorer and CIRI, and users can choose To overcome the problem of rolling circle amplification in reverse
different software for analysis of circRNAs [47,48]. Different circRNAs transcription, Chen et al. developed an assay based on reverse
discovered by these methods require later validation by alternative transcription-droplet digital polymerase chain reaction (RT-ddPCR)
detection methods. This is a key step for the detection of putative [58]. ddPCR is a newly emerging technique for the quantification of
circRNAs by bioinformatic algorithms. nucleic acids with superior sensitivity and accuracy compared with
Unlike microarrays, deep sequencing reveals the actual structure of traditional PCR [59]. The algorithm of this assay is based on Poisson
the circRNA, allows for the discovery of novel circRNAs that have not distribution, which eliminates the influence of rolling RT products on
been previously identified and is not limited by problems with cross circRNA quantification [60]. Similar to a traditional qRT-PCR assay,
hybridization or background noise. In addition, deep-sequencing data total RNA was first isolated and purified from serum, but unlike in
can be generated from formalin-fixed paraffin-embedded (FFPE) speci­ qRT-PCR, it was not treated with RNase R. cDNA was obtained by
mens, making large-scale clinical sample determination possible. How­ reverse transcription and then treated with RNase H to specifically
ever, RNA sequencing efficiency is low, with only 0.1% of reads useable degrade the RNA strand in the DNA-RNA duplex. Subsequently, thermal
for interpretation of a specific splice junction [10,49]. In addition, low cycling was performed using ddPCR. Guo et al. used this assay to profile
abundance circRNAs can be overlooked in common sequencing depths circRNA in plasma and found that the overall survival of gastric cancer
[50]. Fortunately, increased demand for circRNA profiling is driving patients correlates positively with the plasma level of has-circ-0001017
advances in this technology, and deep sequencing analysis for circRNA or has-circ-0061,276 [61]. In addition, patients whose plasma
could be routine in the future with a relatively low price. has-circ-0001017 or has-circ-0061,276 levels returned to normal after
undergoing surgery had a longer disease-free survival than whose did
2.1.4. Fluorescence in situ hybridization (FISH) not return to normal. RT-ddPCR, unlike qRT-PCR, does not require
FISH is a powerful tool that allows the subcellular localization of normalization or calibration steps, thus simplifying the experimental
RNA [51,52]. In this technique, fixed cells or tissues are exposed to short processes. Moreover, its sensitivity makes it suitable for detecting tar­
fluorescently labeled DNA probes in sufficiently high concentrations to gets with low copy numbers without pre-amplification steps [62]. This
form stable DNA-RNA hybrids. After hybridization, fluorescent signals technique requires costly equipment for signal detection. As long as it
are visualized with a fluorescence microscope. DNA probes to different can be accepted by ordinary medical laboratories, this technology is
target molecules, each labeled with a unique fluorophore, can be pooled promising.
before use [52]. Creating probes that target the back-splice junctions of In 2020, Li et al. developed an alternative PCR-based assay that
circRNAs is challenging. Probes should be 22–24 nucleotides to reduce avoids the problems of linear RNA interference and rolling circle
non-specific hybridization, and should complement both sides of the amplification (Fig. 2) [63]. This assay uses a pair of DNA probes
back-splice junction sites [53]. In addition, the GC content of the probe designed to complement each side of the backsplice junction site of a
should fall between 40% and 60% [54]. It is well noted that the target target circRNA. Each DNA probe contains a primer-specific sequence
copy number, probe modification, and conditions of stringency influ­ and a target-specific recognition sequence that complements one side of
ence the signal of in situ hybridization. Validations including proficiency the back-splice junction site of circRNA. In the presence of circRNA, the
testing, assessment of employee competency, instrument calibration, two probes hybridize on either side of the back-splice junction site,
and correlation with clinical findings are required prior to applying allowing their ends to be ligated by splintR ligase [64,65], thereby
FISH-based detection of circRNAs in clinical settings. Additionally, a creating a single long DNA strand. Once the long DNA strand is formed,
more sensitive method is required for differentiating equivocal samples. the primer-specific sequences on either end allow PCR amplification. In
Extending the approach of heavily labeled probes or using an linear RNA, the two probes recognize the border position at its 3ʹ and 5ʹ
enzyme-labeled fluorescence probe may be future direction. However, termini, preventing ligation and subsequent PCR amplification. This
this technique is time-consuming and requires costly equipment for assay can attain a limit of quantitation (LOQ) of 1 fM, with a dynamic
signal detection, which may hinder its utilization in routine diagnostic range of five orders of magnitude [63]. Compared with traditional

3
Z. Mi et al. Talanta 238 (2022) 123066

revealed a detection range of four orders of magnitude, from 2 fM to 50

pM of circRNA. Compared with traditional qRT-PCR, this method has
the advantages of simple operation, low cost, and not requiring for
high-precision temperature cycling and additional separation steps.
The subcellular location of newly discovered circRNAs is of partic­
ular interest. Traditional PCR-based methods can profile circRNA in a
sample but are unable to localize gene expression to morphological
features in tissue sections. Recently, Li et al. established a new assay that
combines RCA with linear DNA nanostructures (LDN) for in situ circRNA
profiling (Fig. 4) [70]. In their study, a long linear DNA scaffold was
produced from a circRNA by RCA which contained multiple binding
sites for two different kinds of hairpin probes. Hairpin-shaped probes
(H1 and H2 probes) hybridized with the DNA scaffold, assembled into
intact LDNs. Once the LDN is formed, the target circRNA hybridizes with
the H1 probe on the LDN, triggering a conformational change in both
probes that allows fluorescent signaling by H2. The target circRNA is
then released, freeing it to react with additional H1 probes along the
scaffold, triggering an enzyme-free signal amplification reaction that
results in the whole LDN lighting up. A single target circRNA can cata­
lyze multiple LDNs, leading to signal amplification. The limit of detec­
tion (LOD) of this assay was approximately 1 pM [70]. However, in their
Fig. 2. Schematic representation of the ligation-based PCR assay for study, the factors influencing probe penetration were not considered.
circRNA detection. Probes are specific for a single target circRNA, and the distances be­
tween the H1 and H2 probes or H1/H2 groups must be carefully designed.
RT-PCR, this method avoids issues with rolling circle amplification and These factors increase the complexity of this method and make it
is easy to operate in routine laboratories owing to its rapidity, simplicity, time-consuming.
and high cost-effectiveness.
2.2.3. Stem-loop primer-induced double-exponential amplification-based
2.2.2. Rolling cycle amplification based circRNA detection circRNA detection
Rolling circle amplification (RCA) is a recently developed technique As circRNA applications continue to evolve, there is an urgent need
that generates long single-stranded DNA or RNA products from a DNA to develop efficient and reliable detection strategies for circRNA
template [66]. This simple and efficient DNA amplification technique profiling in diagnostic tests or therapies. Loop-mediated isothermal
has been widely adopted in biomedical research to establish sensitive amplification (LAMP) is a DNA amplification process with excellent
diagnostic assays for various biomaterials, including DNA and RNA, as specific recognition ability, superior amplification efficiency, and low
well as proteins, small molecules, and cells [67,68]. The circular struc­ background signal [71]. It has been widely applied for the determination
ture of circRNA is an ideal template for initiating the rolling cycle of miRNAs, viruses, and specific DNA sequences [72,73]. This technol­
amplification process, making this technique especially suitable for the ogy, developed in 2000 by Notomi et al. requires a DNA polymerase with
profiling of circRNAs. strand displacement capability and a three primers pairs that bind six
In 2020, Liu et al. designed a reverse transcription rolling circle different regions on the target DNA, ensuring high sequence selectivity
amplification method to selectively amplify target circRNAs (Fig. 3) [74]. The DNA polymerase has a unique property of strand displace­
[69]. This system uses a single complementary DNA primer and a mo­ ment, and the primer pairs binding to six different regions on the target
lecular beacon for signal detection. The DNA primer binds to the junc­ DNA are applied to ensure high sequence selectivity. This technique
tion site on the target circRNA, allowing reverse transcriptase to begin makes use of site-specific sequence extension and nicking to exponen­
rolling cycle amplification. After amplification, the molecular beacon tially generate and release target products. With this method, trace
hybridizes to a site on the reverse amplification DNA strand, leading to amounts of target can generate a large amount of product in a single
the release of the fluorescent signal. A calibration curve of this assay amplification step that does not require modified or labeled DNA probes,

Fig. 3. Schematic representation of the reverse transcription-rolling circle amplification assay for circRNA detection. Reproduced with permission [69]. Copyright
2020. Elsevier.

4
Z. Mi et al. Talanta 238 (2022) 123066

Fig. 4. Schematic illustration of LDN-based method for circRNAs detection. (A) Synthetic procedure to prepare linear DNA nanostructure (LDN); (B) LDN-based
method for circRNAs detection in cells. Reproduced with permission [70]. Copyright 2020. American Chemical Society.

making it a highly specific detection method [75]. reactions that impact testing accuracy. As a result, nuclease-assisted
In 2020, Li et al. developed an ultrasensitive circular RNA detection amplification methods have been developed as an alternative for
assay that uses stem-loop primers (SLP) to induce exponential amplifi­ circRNA profiling.
cation [76]. The SLPs recognize and bind the back-splice junction of In 2018, Li et al. created a strategy to quantify circRNA using its
circRNA, thus forming a double stem-loop DNA structure only in the ability to act as an miRNA sponge and duplex specific nuclease (DSN)
presence of a target circRNA. Once the double stem-loop structure is amplification [80], as shown in Fig. 5. DSN is a nuclease that degrades
created, autocycling strand displacement DNA synthesis can be initiated DNA in double-stranded DNA or DNA:RNA hybrid duplexes but is
by two inner primers. This design directly distinguishes circRNA from incapable of cleaving single-stranded DNA, single-stranded RNA, or
the corresponding linear RNA, precluding the use of RNase R, and avoids double-stranded RNA [81,82]. In their study, molecular beacon probes
overestimation of circRNA due to rolling circle amplification, improving were designed to hybridize with seven binding sites on a target circRNA.
measurement accuracy. The SLP-based technique has very high sensi­ Binding of the molecular beacon to its target induces a conformational
tivity, with an LOQ as low as 10 aM, and has a dynamic range of seven change in its stem-loop structure, triggering the release of a
orders of magnitude, from 10 aM to 100 pM, owing to the incomparable fluorophore-quencher and allowing fluorescence of the probe. DSN then
sensitivity of double exponential amplification. In their proposed assay, digests the DNA strand of the DNA/circRNA hybrid, releasing both the
as low as 10 aM sub-ciRS-7, corresponding to about 60 copies in a re­ fluorescent probe fragment and the circRNA, which can then hybridize
action volume of 10 μL, can be accurately detected. To the best of our with a new molecular beacon probe. This cyclic cleavage of probes by
knowledge, this is the most sensitive method for circRNA profiling. In DSN greatly enhances the fluorescent signal in the presence of the target
their study, the amount of ciRS-7 in 100 ng total RNA extracted from circRNA.
cells was measured as 36.26 aM in K562 cells and 19.03 aM in HT29 Li et al. further simplified circRNA detection, developing an elec­
cells. Indeed, this method is sensitive enough for detection in clinical trochemical assay with back splice junction site recognition and DSN-
samples. Almost all of our current knowledge of the genome has origi­ assisted amplification [83]. In this system, a hairpin probe hybridizes
nated from studies performed at the population level, typically involving with the back splice junction site of a target circRNA. Samples are then
thousands or more of cells analyzed in bulk. The resulting analysis can digested with a DSN that cleaves DNA in long DNA/RNA hybrid strands
provide valuable information; however, it often neglects any heteroge­ (20 bp) but is inactive when hybrid strands are short (<15 bp). Only
neity that occurs within the population of cells [77,78]. The insight into those probes that have bound circRNA, and not linear RNA, can be
the physiologic function of circRNA requires quantifying circRNA cleaved by DSN and trigger recycling amplification, resulting in MB tags
expression at the single-cell level [79]. For single-cell or circulating being removed from the gold disk electrodes surface, finally decreasing
tumor cell analysis, the total size of the sample is much different from current response. This assay can directly detect circRNA in complex
the commonly used size; thus development of more sensitive assays is biological samples at a LOD as low as 3.5 fM and does not require pre­
still in essentially needed. For this assay, the templates and primers are treatment with RNase R. Thus, this electrochemical platform provides a
specific for a single target and require elaborate design as compared simple strategy for profiling circRNA without costly equipment or
with RT-PCR, making the method complicated and costly to use in complicated operations.
clinical applications. DSN-based assays, while amplifying signal in the presence of a target
circRNA, only achieve a 35-fold enhancement in fluorescence as
2.2.4. Enzyme-based circRNA detection compared with a control without DSN. The amplification efficiency of
RCA and LAMP amplification systems have attracted considerable DSN cannot be compared with that of PCR or RCA. Finding or engi­
attention because of their ability to produce a detectable signal from neering new enzymes with better amplification efficiency or combining
trace amounts of target circRNA. However, both methods depend on sets with other isothermal amplification will facilitate the practical appli­
of probe sequences or primers that must be elaborately designed. In cation of the nuclease-assisted amplification methods for circRNA
addition, these detection systems are complicated and may involve side detection.

5
Z. Mi et al. Talanta 238 (2022) 123066

Fig. 5. Schematic representation of the duplex-specific nuclease based assay for circRNA detection.

3. Summary and prospects circRNA expression profiles and specific diseases. Given the potential
use of circRNA as a noninvasive diagnostic and prognostic biomarker,
Over the past decade, numerous studies have demonstrated that accurate profiling of circRNA is an essential prerequisite for clinical
circRNAs are aberrantly expressed in diseased tissues, making them application.
potentially useful in routine clinical diagnosis and therapy. The char­ At present, conventional detection techniques such as qRT-PCR,
acteristics of circRNA such as conservation, abundance, tissue-specific microarray, and RNA-seq are widely employed to profile circRNA.
expression, and stability make them ideal biomarkers. To date, Each technique presents distinct advantages for the quantification,
research has mainly focused on uncovering the relevance between screening, and discovery of circRNAs, as well as limitations such as false

Table 1
List of advantages and disadvantages of available circRNA detection methods.
Detection Mechanism Advantage Limitation Detection Reference
method limit

Traditional cDNA was firstly synthesised by reverse High sensitivity; Quantitative and easy Costly; Femtomolar [21]
qRT-PCR transcription, then PCR amplification was taken taken method; Powerful technique for Low-throughput; range
using divergent primers expression study Elaborated analysis
ddRT-PCR As same for traditional qRT-PCR, and the algorithm Absolute quantification; high sensitivity Costly; Low-throughput Femtomolar [58]
is based on Poisson distribution and accuracy; it eliminates the effect of range
rolling RT products
Ligation- Ligation of two designed DNA probes at the junction High sensitivity and specificity Low-throughput 1 fM [63]
based site of circRNA, followed by PCR of the ligated DNA
PCR probes
RNA Transcriptome-wide profile of circRNAs in any Investigate all circRNAs of any size, known Relatively high costs; Time __ [42,43]
sequence species and unknown; High sensitivity and consuming and complicated
specificity analysis of data
Northern blot Complementary labeled oligonucleotide probe Widely used method; Good specificity Low sensitivity; Require Nanomolar [30]
binds to a target circRNA captured on a large amounts of RNA range
nitrocellulose membrane sample; Time-consuming
Microarray Based on nucleic acid hybridization between target High-throughput circRNA profiling, Test Non-quantitative; Femtomolar [25]
molecules and their corresponding complementary multiple samples simultaneously; Rapid Low sensitivity; range
probe immobilized on to circRNA microarray chips and computationalized analysis Low specificity
FISH Fixed cells or tissues are exposed to short Subcellular localization of circRNA Non-quantitative; __ [52]
fluorescently labeled DNA probes for hybridization Time-consuming;
costly
LDN-based Target circRNA initiates cascade displacement Quantitative detection and imaging of Low-throughput; Time- 1 pM [70]
system reaction between two hairpin probes sequentially circRNA in cells consuming
assembled along the LDN
RT-RCA using primer binds to the junction site on the target Simple operation; Low cost; High Low-throughput 2 fM [69]
circRNA, allowing reverse transcriptase to begin selectivity
rolling cycle amplification
LAMP SLPs bind the back-splice junction to form a double High sensitivity; High specificity Elaborated design; Costly; 10 aM [76]
stem-loop DNA structure, followed by autocycling Low-throughput
strand displacement DNA synthesis

6
Z. Mi et al. Talanta 238 (2022) 123066

positive results, long processing time, high intensity labor demands, of multiple circRNAs. Thus, a high-throughput method for simultaneous
sample size requirements, and low efficiency. Therefore, significant ef­ circRNA detection is of great significance. An integrated microfluidic
forts have been made to improve circRNA profiling, resulting in the chip may be one choice for providing a powerful tool for high-
development of a variety of new approaches, such as RCA and LAMP. In throughput analysis of circRNA in single-cell gene expression measure­
this review, we provided insights into these new methods; the advan­ ments. (iii) The LOD of the most of the current assays is on the fmol
tages and limitations are summarized in Table 1. Northern blotting is scale, and pursuing ultrasensitive methods is essential for profiling low-
currently considered the gold standard for confirming the presence of abundance circRNA, especially for single-cell or circulating tumor cell
circRNAs; however, it is limited by low sensitivity, low throughput, and circRNA analysis. One direction involves the use of nanomaterials, such
time-consuming processes. Strategies for increasing the sensitivity and as metallic nanoparticles or semiconductor quantum dots. The nano­
reducing the total assay time may improve its practicality. Microarray materials can produce a synergic effect among catalytic activity, con­
analysis is a powerful tool for the identification of differentially ductivity, and biocompatibility to accelerate the signal transduction,
expressed circRNAs at a large-scale in parallel, but, it requires further leading to the ultrasensitive detection by miniaturized devices. Novel
validation to quantify the expression more accurately. Determining how physical biosensors employing total internal reflection fluorescence
to decease the background signal as well as cross-hybridization should microscopy, Raman scattering and surface plasmon resonance are sup­
be considered for improving the accuracy. Deep sequencing allows for posed to be an alternative technique for circRNA profiling owing to the
the discovery of novel circRNAs that have not been previously identified unique optical, electronic, and magnetic properties of nanomaterials.
via cloning or sequencing and is not subject to the problems of back­ (iv) Technologies for direct detection of the temporal and spatial
ground noise and cross hybridization. The industry is now focusing on expression sequence of specific circRNA in tissue patterns are extremely
this technology, and in the future, when the testing price becomes important for elucidating circRNA biology. Because of the low level of
acceptable, it can be routinely used in clinical testing. FISH is a powerful circRNA in the cells, it is essential to further develop a more efficient and
technique for visualization of the spatial localization of circRNA at tis­ sensitive circRNA detection method in vivo. The development of new
sue, cellular, and even subcellular levels. Validation, including profi­ profiling techniques would provide scientists with more choices for the
ciency testing, assessment of employee competency, instrument quantitative detection of circRNA. From a practical point of view, we
calibration, and correlation with clinical findings is required before its it should consider the advantages and limitations of each approach to
can be used in clinical practice. qRT-PCR is the most used assay for obtain the most economical and efficient experimental method. In the
circRNA profiling; it represents a balance of sensitivity, specificity, future, novel approaches in circRNA profiling will present more
speed, cost, and simplicity, which makes this technique easy to carry out powerful tools for improving human health in early diagnostics.
in a clinical laboratory. However, it is subject to strand displacement
and rolling circle replication during reverse transcription, leading to an
overestimation of circRNA level. RT-ddPCR is an alternative technology Declaration of competing interest
that eliminates the influence of rolling RT products, with superior
sensitivity and accuracy compared with traditional PCR, and it holds The authors declare that they have no known competing financial
great promise for use in the clinical practice. Other amplification-based interests or personal relationships that could have appeared to influence
methods, such as ligation-based PCR and RT-RCA assays, are easy to the work reported in this paper.
operate in routine laboratories owing to their rapidity, simplicity, and
high cost-effectiveness. More samples or laboratories are needed to Acknowledgments
verify its usability in clinical testing. The stem-loop primer induced
amplification based assay is currently the most sensitive method for This study was supported by the National Natural Science Founda­
circRNA profiling; the templates and primers require elaborate design, tion of China (grant no.82002252).
making it complicated and costly to use in clinical applications.
Reducing the complexity of primer design while preserving the sensi­
References
tivity may increase its potential for clinical use. A nuclease-assisted
amplification assay, such as DSN, is an alternative assay for circRNA [1] F.R. Kulcheski, A.P. Christoff, R. Margis, Circular RNAs are miRNA sponges and can
profiling. However, limited amplification efficiency precludes its use in be used as a new class of biomarker, J. Biotechnol. 238 (2016) 42–51.
routine clinical diagnostics. Finding more efficient enzymes or [2] L.S. Kristensen, M.S. Andersen, L. Stagsted, K.K. Ebbesen, J. Kj Em S, The
biogenesis, biology and characterization of circular RNAs, Nat. Rev. Genet. 20
combining with other amplification reactions may result in a significant (2019) 675–691.
breakthrough. Taking accuracy, sensitivity, speed, cost, operation dif­ [3] Y. Liang, N. Liu, L. Yang, J. Tang, M. Mei, A brief review of circRNA biogenesis,
ficulty and instrumentation required into account, in our opinion, detection and function, Curr. Genom. 22 (2021) 1–11.
[4] W.R. Jeck, J.A. Sorrentino, K. Wang, M.K. Slevin, C.E. Burd, J. Liu, W.F. Marzluff,
ddPCR- and RCA-based assays may be the best choice for clinical diag­ N.E. Sharpless, Circular RNAs are abundant, conserved, and associated with ALU
nosis currently. repeats, RNA 19 (2013) 141–157.
Numerous strategies for circRNA profiling have been described, but [5] J. Li, D. Sun, W. Pu, J. Wang, Y. Peng, Circular RNAs in cancer: biogenesis,
function, and clinical significance, Trends Cancer 6 (2020) 319–336.
some obstacles still exit. (i) In general, circRNA profiling in vitro requires [6] J. Salzman, C. Gawad, P.L. Wang, N. Lacayo, P.O. Brown, Circular RNAs are the
RNA isolation and separation from linear RNA, commonly by treatment predominant transcript isoform from hundreds of human genes in diverse cell
of RNA samples with an exonuclease enzyme (commonly RNAse R). types, PLoS One 7 (2012), e30733.
[7] Q. Vicens, E. Westhof, Biogenesis of circular RNAs, Cell 159 (2014) 13–14.
Thus, linear RNAs are processed leaving the circRNAs intact. However, it
[8] X.-O. Zhang, H.-B. Wang, Y. Zhang, X. Lu, L.-L. Chen, L. Yang, Complementary
has recently been indicated that some circRNAs are sensitive to this sequence-mediated exon circularization, Cell 159 (2014) 134–147.
enzyme. This treatment could interfere with the global selection of the [9] T.B. Hansen, T.I. Jensen, B.H. Clausen, J.B. Bramsen, B. Finsen, C.K. Damgaard,
J. Kjems, Natural RNA circles function as efficient microRNA sponges, Nature 495
circRNAs within a sample, producing a bias. Employment of additional
(2013) 384–388.
biochemical procedures for the isolation of circRNAs, such as two- [10] X. Yang, T. Ye, H. Liu, P. Lv, C. Duan, X. Wu, K. Jiang, H. Lu, D. Xia, E. Peng,
dimensional denaturing polyacrylamide gel electrophoresis, rRNA Z. Chen, K. Tang, Z. Ye, Expression profiles, biological functions and clinical
depletion, and poly(A)-depletion, to increase the amount of circRNAs in significance of circRNAs in bladder cancer, Mol. Cancer 20 (2021) 1–25.
[11] T. Xu, J. Wu, P. Han, Z. Zhao, X. Song, Circular RNA expression profiles and
samples has been proposed, but these treatments are time-consuming. features in human tissues: a study using RNA-seq data, BMC Genom. 18 (2017)
There is a great need to further develop a simple, low cost, and high- 131–142.
efficiency pretreatment assay for circRNA isolation or separation. (ii) [12] S.P. Barrett, J. Salzman, Circular RNAs: analysis, expression and potential
functions, Development 143 (2016) 1838–1847.
Commonly, one circRNA can regulate many functions and is relevant to [13] N. Abu, R. Jamal, Circular RNAs as promising biomarkers: a mini-review, Front.
a number of diseases. One disease is often associated with the expression Physiol. 7 (2016) 355.

7
Z. Mi et al. Talanta 238 (2022) 123066

[14] S. Chaichian, R. Shafabakhsh, S.M. Mirhashemi, B. Moazzami, Z. Asemi, Circular [47] X. Zhang, H. Wang, Y. Zhang, X. Lu, L. Chen, L. Yang, Complementary sequence-
RNAs: a novel biomarker for cervical cancer, J. Cell. Physiol. 235 (2020) 718–724. mediated exon circularization, J. Cell. 159 (2014) 134–147.
[15] H.D. Zhang, L.H. Jiang, D.W. Sun, J.C. Hou, Z.L. Ji, CircRNA: a novel type of [48] Y. Gao, J. Wang, F. Zhao, CIRI: an efficient and unbiased algorithm for de novo
biomarker for cancer, Breast Cancer 25 (2018) 1–7. circular RNA identification, Genome Biol. 16 (2015) 4.
[16] Z. Zhang, T. Yang, J. Xiao, Circular RNAs: promising biomarkers for human [49] Y. Zhang, X.O. Zhang, T. Chen, J.F. Xiang, Q.F. Yin, Y.H. Xing, S. Zhu, Y. Li, L.
diseases, EBioMedicine 34 (2018) 267–274. L. Chen, Circular intronic long noncoding RNAs, Mol. Cell. 51 (2013) 792–806.
[17] H. Tabak, G. Van der Horst, J. Smit, A. Winter, Y. Mul, G.M. Koerkamp, [50] Z. Li, C. Huang, C. Bao, L. Chen, M. Lin, X. Wang, G. Zhong, B. Yu, W. Hu, L. Dai,
Discrimination between RNA circles, interlocked RNA circles and lariats using two- P. Zhu, Z. Chang, Q. Wu, Y. Zhao, Y. Jia, P. Xu, H. Liu, G. Shan, Exon-intron
dimensional polyacrylamide gel electrophoresis, Nucleic Acids Res. 16 (1988) circular RNAs regulate transcription in the nucleus, Nat. Struct. Mol. Biol. 22
6597–6605. (2015) 256–264.
[18] W.R. Jeck, N.E. Sharpless, Detecting and characterizing circular RNAs, Nat. [51] N.M. Chrzanowska, J. Kowalewski, M.A. Lewandowska, Use of fluorescence in situ
Biotechnol. 32 (2014) 453–461. hybridization (FISH) in diagnosis and tailored therapies in solid tumors, Molecules
[19] E.W. Khandjian, C. Méric, A procedure for Northern blot analysis of native RNA[J], 25 (2020) 1864.
Anal. Biochem. 159 (1986) 227–232. [52] A.S. Lim, T.H. Lim, Fluorescence in situ hybridization on tissue sections, Methods
[20] J. Cocquet, A. Chong, G. Zhang, R.A. Veitia, Reverse transcriptase template Mol. Biol. 1541 (2017) 119–125.
switching and false alternative transcripts, Genomics 88 (2006) 127–131. [53] S. Memczak, M. Jens, A. Elefsinioti, F. Torti, J. Krueger, A. Rybak, L. Maier, S.
[21] C.-Y. Yu, H.-J. Liu, L.-Y. Hung, H.-C. Kuo, T.-J. Chuang, Is an observed non-co- D. Mackowiak, L.H. Gregersen, M. Munschauer, A. Loewer, U. Ziebold,
linear RNA product spliced in trans, in cis or just in vitro? Nucleic Acids Res. 42 M. Landthaler, C. Kocks, F. le Noble, N. Rajewsky, Circular RNAs are a large class
(2014) 9410–9423. of animal RNAs with regulatory potency, Nature 495 (2013) 333–338.
[22] S.P. Barrett, P.L. Wang, J. Salzman, Circular RNA biogenesis can proceed through [54] Y. Wada, Y. Ohta, M. Xu, S. Tsutsumi, T. Minami, K. Inoue, D. Komura, J.
an exon-containing lariat precursor, Elife 4 (2015), e07540. i. Kitakami, N. Oshida, A. Papantonis, A wave of nascent transcription on activated
[23] D.A. Cooper, M. Cortés-López, P. Miura, Genome-wide circRNA profiling from human genes, Proc. Natl. Acad. Sci. U.S.A. 106 (2009) 18357–18361.
RNA-seq data, Methods Mol. Biol. 1724 (2018) 27–41. [55] A.C. Panda, M. Gorospe, Detection and analysis of circular RNAs by RT-PCR, Bio.
[24] Y. Li, J. Zhao, S. Yu, Z. Wang, X. He, Y. Su, T. Guo, H. Sheng, J. Chen, Q. Zheng, Protoc. 8 (2018), e2775.
Extracellular vesicles long RNA sequencing reveals abundant mRNA, circRNA, and [56] M. Jinek, J.A. Doudna, A three-dimensional view of the molecular machinery of
lncRNA in human blood as potential biomarkers for cancer diagnosis, Clin. Chem. RNA interference, Nature 457 (2009) 405–412.
265 (2019) 798–808. [57] H. Suzuki, T. Tsukahara, A view of pre-mRNA splicing from RNase R resistant
[25] S. Li, S. Teng, J. Xu, G. Su, Y. Zhang, J. Zhao, S. Zhang, H. Wang, W. Qin, Z.J. Lu, RNAs, Int. J. Mol. Sci. 15 (2014) 9331–9342.
Microarray is an efficient tool for circRNA profiling, Briefings Bioinf. 20 (2019) [58] D.-F. Chen, L.-J. Zhang, K. Tan, Q. Jing, Application of droplet digital PCR in
1420–1433. quantitative detection of the cell-free circulating circRNAs, Biotechnol. Biotechnol.
[26] Y. Xi, M. Fowdur, Y. Liu, H. Wu, M. He, J. Zhao, Differential expression and Equip. 32 (2018) 116–123.
bioinformatics analysis of circRNA in osteosarcoma, Biosci. Rep. 39 (2019). [59] R.H. Sedlak, J. Kuypers, K.R. Jerome, A multiplexed droplet digital PCR assay
BSR20181514. performs better than qPCR on inhibition prone samples, Diagn. Microbiol. Infect.
[27] J. Mao, T. Li, D. Fan, H. Zhou, X. Wang, Abnormal expression of rno_circRNA_ Dis. 80 (2014) 285–286.
014900 and rno_circRNA_005442 induced by ketamine in the rat hippocampus, [60] Y. Uchiyama, M. Nakashima, S. Watanabe, M. Miyajima, M. Taguri, S. Miyatake,
BMC Psychiatr. 20 (2020) 1–12. N. Miyake, H. Saitsu, H. Mishima, A. Kinoshita, Ultra-sensitive droplet digital PCR
[28] M. Wang, P. Su, Y. Liu, X. Zhang, S. Gu, Abnormal expression of circRNA_089763 for detecting a low-prevalence somatic GNAQ mutation in Sturge-Weber syndrome,
in the plasma exosomes of patients with postoperative cognitive dysfunction after Sci. Rep. 6 (2016) 22985.
coronary artery bypass grafting, Mol. Med. Rep. 20 (2019) 2549–2562. [61] T. Li, Y. Shao, L. Fu, Y. Xie, L. Zhu, W. Sun, R. Yu, B. Xiao, J. Guo, Plasma circular
[29] J. Salzman, Circular RNA expression: its potential regulation and function, Trends RNA profiling of patients with gastric cancer and their droplet digital RT-PCR
Genet. 32 (2016) 309–316. detection, J. Mol. Med. 96 (2018) 85–96.
[30] T. Schneider, S. Schreiner, C. Preuer, A. Bindereif, O. Rossbach, Northern blot [62] H. Li, R. Bai, Z. Zhao, L. Tao, M. Ma, Z. Ji, M. Jian, Z. Ding, X. Dai, F. Bao, A. Liu,
analysis of circular RNAs, Methods Mol. Biol. 1724 (2018) 119–133. Application of droplet digital PCR to detect the pathogens of infectious diseases,
[31] W.V. Burnett, Northern blotting of RNA denatured in glyoxal without buffer Biosci. Rep. 38 (2018). BSR 20181170.
recirculation, Biotechniques 22 (1997) 668–671. [63] P. Zhang, N. Guo, K. Gao, F. Su, F. Wang, Z. Li, Direct recognition and sensitive
[32] E. D’Ambra, M. Morlando, Study of circular RNA expression by nonradioactive detection of circular RNA with ligation-based PCR, Org. Biomol. Chem. 18 (2020)
northern blot procedure, Methods Mol. Biol. 2348 (2021) 371–383. 3269–3273.
[33] X. Wang, S. Ge, Nonradioactive northern blot of circRNAs, Methods Mol. Biol. 1724 [64] G. Lohman, Y. Zhang, A.M. Zhelkovsky, E.J. Cantor, T.C. Evans, Efficient DNA
(2018) 135–141. ligation in DNA–RNA hybrid helices by chlorella virus DNA ligase, Nucleic Acids
[34] É. Várallyay, J. Burgyán, Z. Havelda, MicroRNA detection by northern blotting Res. 42 (2014) 1831–1844.
using locked nucleic acid probes, Nat. Protoc. 3 (2008) 190–196. [65] H. Wang, H. Wang, C. Liu, X. Duan, Z. Li, Ultrasensitive detection of telomerase
[35] K. Damm, S. Bach, K. Müller, G. Klug, R.K. Hartmann, Improved northern blot activity in a single cell using stem-loop primer-mediated exponential amplification
detection of small RNAs using EDC crosslinking and DNA/LNA probes, Methods (SPEA) with near zero nonspecific signal, Chem. Sci. 7 (2016) 4945–4950.
Mol. Biol. 1296 (2015) 41–51. [66] N.I. Goo, D.E. Kim, Rolling circle amplification as isothermal gene amplification in
[36] W. Tang, K. Fu, H. Sun, D. Rong, H. Wang, H. Cao, CircRNA microarray profiling molecular diagnostics, Biochip. J. 10 (2016) 262–271.
identifies a novel circulating biomarker for detection of gastric cancer, Mol. Cancer [67] M.M. Ali, F. Li, Z. Zhang, K. Zhang, D.-K. Kang, J.A. Ankrum, X.C. Le, W. Zhao,
17 (2018) 1–6. Rolling circle amplification: a versatile tool for chemical biology, materials science
[37] Z. Wang, P. Xu, B. Chen, Z. Zhang, W. Peng, Identifying circRNA-associated-ceRNA and medicine, Chem. Soc. Rev. 43 (2014) 3324–3341.
networks in the hippocampus of Aβ1-42-induced Alzheimer’s disease-like rats [68] C. Yao, R. Zhang, J. Tang, D. Yang, Rolling circle amplification (RCA)-based DNA
using microarray analysis, Aging 10 (2018) 775–788. hydrogel, Nat. Protoc. (2021) 1–24.
[38] Y.F. Zhou, L.J. Shi, J. Yao, Y.N. Sun, B. Yan, Microarray analysis of circRNA [69] Y. Liu, X. Zhang, M. Liu, F. Xu, Q. Zhang, Y. Zhang, X. Weng, S. Liu, Y. Du, X. Zhou,
expression pattern in corneal neovascularization, Cornea 38 (2019) 1443–1449. Direct detection of circRNA in real samples using reverse transcription-rolling
[39] W. Liu, J. Zhang, C. Zou, X. Xie, Y. Wang, B. Wang, Z. Zhao, J. Tu, X. Wang, H. Li, circle amplification, Anal. Chim. Acta 1101 (2020) 169–175.
Microarray expression profile and functional analysis of circular RNAs in [70] J. Jiao, Y. Xiang, C. Duan, Y. Liu, G. Li, Lighting up circRNA using a linear DNA
osteosarcoma, Cell. Physiol. Biochem. 43 (2017) 969–985. nanostructure, Anal. Chem. 92 (2020) 12394–12399.
[40] Q. Wu, Y. Wang, M. Cao, V. Pantaleo, J. Burgyan, W.-X. Li, S.-W. Ding, Homology- [71] T. Notomi, Y. Mori, N. Tomita, H. Kanda, Loop-mediated isothermal amplification
independent discovery of replicating pathogenic circular RNAs by deep sequencing (LAMP): principle, features, and future prospects, J. Microbiol. 53 (2015) 1–5.
and a new computational algorithm, Proc. Natl. Acad. Sci. U.S.A. 109 (2012) [72] K. Nagamine, T. Hase, T. Notomi, Accelerated reaction by loop-mediated
3938–3943. isothermal amplification using loop primers, Mol. Cell. Probes 16 (2002) 223–229.
[41] R. Xin, Y. Gao, Y. Gao, R. Wang, K.E. Kadash-Edmondson, B. Liu, Y. Wang, L. Lin, [73] N. Tomita, Y. Mori, H. Kanda, T. Notomi, Loop-mediated isothermal amplification
Y. Xing, isoCirc catalogs full-length circular RNA isoforms in human (LAMP) of gene sequences and simple visual detection of products, Nat. Protoc. 3
transcriptomes, Nat. Commun. 12 (2021) 1–11. (2008) 877–882.
[42] J.O. Westholm, P. Miura, S. Olson, S. Shenker, B. Joseph, P. Sanfilippo, S. [74] T. Notomi, H. Okayama, H. Masubuchi, T. Yonekawa, K. Watanabe, N. Amino,
E. Celniker, B.R. Graveley, E.C. Lai, Genome-wide analysis of drosophila circular T. Hase, Loop-mediated isothermal amplification of DNA, Nucleic Acids Res. 28
RNAs reveals their structural and sequence properties and age-dependent neural (2000) e63.
accumulation, Cell Rep. 9 (2014) 1966–1980. [75] Y.P. Wong, S. Othman, Y.L. Lau, S. Radu, H.Y. Chee, Loop-mediated isothermal
[43] M. Danan, S. Schwartz, S. Edelheit, R. Sorek, Transcriptome-wide discovery of amplification (LAMP): a versatile technique for detection of micro-organisms,
circular RNAs in Archaea, Nucleic Acids Res. 40 (2012) 3131–3142. J. Appl. Microbiol. 124 (2018) 626–643.
[44] A. Fonseca, Rung Nuno, Brazma Johan, Marioni Alvis, C. John, Tools for mapping [76] P. Zhang, K. Gao, Y. Liang, F. Su, F. Wang, Z. Li, Ultrasensitive detection of circular
high-throughput sequencing data, Bioinformatics 28 (2012) 3169–3177. RNA by accurate recognition of the specific junction site using stem-loop primer
[45] L. Chen, Y. Yu, X. Zhang, C. Liu, C. Ye, L. Fan, PcircRNA_finder: a software for induced double exponential amplification, Talanta 217 (2020) 121021.
circRNA prediction in plants, Bioinformatics 32 (2016) 3528–3529. [77] C. Paolillo, E. Londin, P. Fortina, Single-cell genomics, Clin. Chem. 65 (2019)
[46] M. Niu, J. Zhang, Y. Li, C. Wang, Z. Liu, H. Ding, Q. Zou, Q. Ma, CirRNAPL: a web 972–985.
server for the identification of circRNA based on extreme learning machine, [78] B. Lim, Y. Lin, N. Navin, Advancing cancer research and medicine with single-cell
Comput. Struct. Biotec. 18 (2020) 834–842. genomics, Cancer Cell 37 (2020) 456–470.

8
Z. Mi et al. Talanta 238 (2022) 123066

[79] S. Nisar, A.A. Bhat, M. Singh, T. Karedath, A. Rizwan, S. Hashem, P. Bagga, characterization and molecular cloning of duplex-specific nuclease from the
R. Reddy, F. Jamal, S. Uddin, G. Chand, D. Bedognetti, W. El-Rifai, M.P. Frenneaux, hepatopancreas of the Kamchatka crab, BMC Biochem. 9 (2008) 1–12.
M.A. Macha, I. Ahmed, M. Haris, Insights into the role of circRNAs: biogenesis, [82] Y. Hirakawa, I. Ohiso, Cloning, Expression, and Characterization of Thermostable
characterization, functional, and clinical impact in human malignancies, Front. Duplex-specific Nuclease from Snow Crab (Chionoecetes Opilio), The 32nd Annual
Cell Dev. Biol. 9 (2021), 0617281. Meeting of the Molecular Biology of Japan, Conference, 2009.
[80] J. Jiao, T. Gao, H. Shi, A. Sheng, Y. Xiang, Y. Shu, G. Li, A method to directly assay [83] J. Jin, L. Chao, L. Ning, L. Shi, L. Wang, Y. Xiang, G. Li, Electrochemical detection
circRNA in real samples, Chem. Commun. 54 (2018) 13451–13454. of circRNAs based on the combination of back-splice junction and duplex-specific
[81] V.E. Anisimova, D.V. Rebrikov, D.A. Shagin, V.B. Kozhemyako, N.I. Menzorova, D. nuclease, Sensor. Actuator. B Chem. 302 (2020) 127166.
B. Staroverov, R. Ziganshin, L.L. Vagner, V.A. Rasskazov, S.A. Lukyanov, Isolation,