Documentos de Académico
Documentos de Profesional
Documentos de Cultura
Review
a r t i c l e
i n f o
Article history:
Received 24 March 2016
Received in revised form 21 June 2016
Accepted 15 July 2016
Available online 19 July 2016
Keywords:
Circular RNAs
MicroRNA sponges
IRES sequences
CircRNA-derived protein
Translation
a b s t r a c t
Circular RNAs (circRNAs) are a new class of long non-coding RNAs that play a potential role in gene expression
regulation, acting as efcient microRNAs sponges. The latest surprise concerning circRNAs is that we now
know that they can serve as transcriptional activators in human cells, indicating that circRNAs are involved in important regulatory tasks. Recently, new insight has been gained about the coding potential of circular viroid RNAs,
as well as the presence of Internal Ribosomal Entry Sites (IRES) allowing the formation of peptides or proteins
from circular RNA. Here, we discuss the current state of our knowledge regarding evidence supporting the hypothesis that circRNAs serve as protein-coding sequences in vitro and in vivo. Also, we remark on the difculties
of their identication and highlight some tools currently available for exploring the coding potential of circRNA.
2016 Elsevier B.V. All rights reserved.
1. Introduction
Alternative RNA splicing represents an additional pathway to increase the complexity of the transcriptome by modulating the sequential arrangement of exons, in which particular exons of a gene may be
included or excluded from the nal processed messenger RNA
(mRNA) [1]. The versatility of RNA appears unlimited, and among the
variety of tasks performed by RNA, it can form circular molecules. In
the late-seventies, plant viroids were the rst described naturally
occurring circular RNA molecules [2,3]. Since then, several mammalian
species of circular RNA molecules have been discovered [46]. The
circular RNA molecules derived from lariat structures (ciRNAs), a
by-product of intron removal during splicing, are formed through
end-joining by a 2-5 phosphodiester bond, which can be cleaved by
the debranching enzyme DBR1, making the molecule available to
degradation by exonucleases [7]. CiRNAs that fail to be debranched can
accumulate in the nucleus, and some have displayed a cis-regulatory
role on their parent coding genes [8].
In contrast, another class of circular RNA molecules, the circRNAs,
comprise an abundant class of highly stable RNAs that consist of 24
exons of protein-coding genes, although these can also derive from
non-coding exons, intergenic regions or transcripts antisense to 5 and
3 UnTranslated Regions (UTRs) [911]. Additionally, circRNAs are
widely expressed in a complex tissue-, cell-type- or developmentalstage-specic manner, and their expression levels can be approximately
10-fold compared with their linear isoform, suggesting that their
Corresponding author at: Dr. Mrquez No. 162, Col. Doctores, Delegacin:
Cuauhtmoc, Mxico D.F. C.P 06720, Mexico.
E-mail address: guillaqui@himfg.edu.mx (G. Aquino-Jarquin).
http://dx.doi.org/10.1016/j.bbagrm.2016.07.009
1874-9399/ 2016 Elsevier B.V. All rights reserved.
formation may be nely regulated [12,13]. Canonical eukaryotic premRNA splicing is catalyzed by the spliceosomal machinery to remove
introns and join exons, leading to the formation of a linear RNA
transcript with 5 to 3 polarity [14]. The biogenesis of circRNAs usually
involves the transcription and splicing of the parent gene, and it has
been suggested that their formation may compete with linear splicing
[15]. Sequence annotation suggests that a subset of circRNAs results
from the direct ligation of the 5 and 3 ends of linear RNA molecules
[16]. They can also originate from intermediates in RNA-processing
reactions, or be the result of the back-splicing of protein-coding
genes, whereby a 5 splice site (the splice donor) is joined to a 3 splice
site (the splice acceptor) located upstream in the same pre-mRNA
molecule [16] (Fig. 1).
Protein-coding regions have been dened according to simple rules
concerning the nature of translation, for example, the presence of Open
Reading Frames (ORFs) with minimal length, practical use of codonusage biases, and the use of AUG as the initiation codon [17]. Recently,
new insight has been gained about the coding potential of circular viroid
RNAs can exhibit coding potential [18], as well as the presence of
Internal Ribosomal Entry Sites (IRES) sequences, which might allow
the formation of peptides or proteins from circular RNA [19,20].
However, because no current substantial evidence is available about
endogenous circRNA-derived protein in eukarya, it has been generally
accepted that circRNAs have a tendency to function as a new class of
long non-coding RNAs (lncRNAs). For now, there is limited knowledge
related to the biological roles of circRNAs in translation; nonetheless,
at present, the possibility that some circRNAs can function as proteincoding RNA transcripts cannot be discarded. Here, our aim was to
highlight some relevant aspects in circRNAs biology and, particularly,
we discuss the possible scenarios by which circRNAs can serve as
1246
Fig. 1. Backsplicing for circRNA formation. Different classes of circular RNAs (circRNAs) can be originated from a particular gene. These circRNAs are generated through a non-canonical
splicing process known as backsplicing in which a downstream splice donor is joined to an upstream splice acceptor. Such circRNAs can be formed by one or more exons and can even
contain unspliced intronic sequences (exon-intron(s)-exon intermediate). ss, splice site [67].
circRNA derivative from the C2H2 zinc nger 91 locus (circRNAZNF91) contains 24 conserved miR-23 sites, surpassing circSry, which
has 16 sites for miR-138. However, even with 24 binding sites,
circRNA-ZNF-91 is far from resembling CDR1as, which contains 71
miR-7 sites, based on the analysis of miRNA site number performed
[24]. This allowed the authors to conclude that no other human circRNA
shows the strong potential capability of CDR1as as a sponge for any
conserved miRNA seed family [24], which denotes that we are only
beginning to understand the magnitude of the biological signicance
of circRNAs.
3. Role of circular RNAs in transcription regulation: more than
sponges
Important advances have been made in identifying and understanding the interconnections within and among the different processes that
regulate gene expression [25]. For many years, it has been known that
small nuclear RNA (snRNAs) and nucleolar RNAs exert their functions
through direct RNARNA interactions [16]. In this regard, one of the
rst evidences on this interaction was from Kwek et al., who reported
that TFIIH specically associates with U1 snRNA in the coordinated
and efcient control of transcriptional initiation and in early mRNA processing [26]. The discovery of the competing endogenous RNA interplay
of circRNAs has increased the likelihood that these circular RNAs may
have an integral role in regulatory RNA networks [27]. Recently, Li
et al. established that circRNAs containing retained intronic sequences
can function as positive regulators of the RNA Pol II transcription of
their parental genes in a U1 snRNA-dependent manner in HEK-293
human cells [28] (Fig. 2, left panel). The authors identied 111 Pol IIinteracting circRNAs, and only 15 were conrmed to contain introns
among the circularized exons, which were referred as ExonIntron
circular RNAs (EIciRNAs). Li et al. demonstrated that EIciEIF3J and
EIciPAIP2, the two most abundant nuclear species of EIciRNAs in
HEK-293, are enriched at transcription sites and may promote the
transcription of their parent mRNAs. These results provide evidence
for cross-talk between initiation of transcription and mRNA splicing
and a novel circRNAs-based synergistic approach for transcriptional
control in a highly orchestrated manner [28].
Although it is not well known what determines which circRNAs can
shuttle between nucleus and cytoplasm, this dynamic exchange might
be facilitated by nuclear export sequences or nuclear localization
sequences, positioned in exons or retained introns, similar to the
nuclear localization of some miRNAs, e.g., miR-29b [29]. Alternatively,
they might be exported by mRNA export factors [30]. Thus, in addition,
1247
Fig. 2. Diagrammatic depiction of circRNAs in transcriptional regulation and their protein-coding potential. A) Positive feedback loop caused by crosstalk between U1 snRNA-EIciRNAs
interactions. In the nucleus of eukaryotic cells, EIciRNAs (e.g., EIciEIF3J and EIciPAIP2) interact with U1 snRNP, RNA Pol II transcription complex and promoter regions to modulate the
expression of the parental genes in cis, thus augmenting the levels of both circRNA and mRNA. B) Translation directed by circular RNA showing the sequence features within a
transcript's ORF that might be involved in protein synthesis, such as IRES elements, RBPs, AUG initiator codon, and stop codon (in some cases). TSS, Transcription Start Site; TFIID, IID;
TFIIA, IIA; TFIIB, IIB; TFIIE, IIE, TFIIF, IIF; U1 RNP, U1 small nuclear ribonucleoprotein; circRNA, circular RNA; mRNA, messenger RNA; ORF, open reading frame; RBPs, RNA-binding
proteins; IRES, internal ribosomal entry site.
1248
Table 1
CircRNAs with experimental data supporting the potential coding in vitro and in vivo.
Model
Virus
Hepatitis
delta ()
virus
(HDV)
Rice yellow
mottle
virus
Bacteria
Escherichia
coli
CircRNA feature
Sequence feature
Peptide/protein
Source
Function
Ref.
Circular
single-stranded
RNA
Protein of 122
amino acids
Exogenous/infectious
Unknown
[33]
Covalently
closed circular
RNA (220 nt)
Ectopic, plasmid-encoded
Unknown
[18]
795-nt circular
mRNA
GFP
Ectopic plasmid-encoded
[34]
IRES-dependent sequence
GFP
Ectopic plasmid-encoded
[19]
NH2-terminal
portion of NCX1
protein (70-kDa)
EGF, IGF-1, IGF-2
Endogenous
[35]
In-vitro
transcription/translation
process, transfection of
synthetic circRNA
[20]
Mammals
HEK-293 cells Single exon
minigene
HEK-293 cells Exonic
Rabbit
Exonic
reticulocyte
lysate
HeLa cells
Exonic
Cap-independent translation,
IRES-independent sequence, poly-A
tail-independent translation
CircRNA, Circular RNA; nt, nucleotide; ORF, open reading frame; GFP, green uorescent protein; IRES, internal ribosomal entry sites; EGF, epidermal growth factor; NCX1, sodium-calcium
exchanger; IGF-1, insulin-like growth factor 1; insulin-like growth factor 2.
that bacterial ribosomes can repeatedly direct the GFP expression from
a circular mRNA [34] (Table 1). In this case, a circular mRNA of 795nucleotides in length, with an innite 30-kDa-encoding ORF, produced
proteins N 300 kDa in size, indicating that the bacterial ribosomes
scanned the circle N10 times [34]. Similarly, Wang et al. evidenced
that a single exon minigene containing split GFP can be efciently
backspliced to generate circRNA and that it can direct the translation
of an entire functional GFP protein inside HEK-293 human cells [19]
(Table 1). Although synthetic circRNAs containing IRES elements can
be translated, currently no concrete evidence is available to suggest
that endogenous circRNAs in eukaryotic cells, containing or not IRES
elements, might encode proteins with functions distinct from those of
their canonical linear counterparts [39].
Thus, the fact that synthetic circRNAs can be efciently translated
in vitro supports the notion that some circRNAs can be translatable molecules in vivo (Fig. 2). Previously, Li et al. described a truncated protein
originating from a circRNA of the Na+/Ca2+ exchanger gene 1 (NCX1)
into HEK-293 transfected cells. When a circular NCX1 exon 2 transcript
was overexpressed ectopically, the size of the truncated protein was
consistent with the predicted molecular weight of approximately
70 kDa, and surprisingly, this protein exhibited Na+/Ca2 + exchange
activity, denoting its function [35]. However, Li et al. could not identify
the same-sized proteins in native tissue, observing a prominent and
slightly smaller band, possibly as a result of the hydrolysis residual of
the circRNA-derived protein, coupled with discrepancies in terms of
antibody detection (length and antigenicity) [35]. Perhaps this is the
only putative example of an endogenous peptide derived from circRNA
identied so far in mammals; however, these inconsistencies do not
support entirely the existence of endogenous circRNA-directed
translation in vivo.
Although it can be assumed that linear lncRNAs are not generally
translated into proteins, growing evidence shows that a subset of
these can be a source of functional small or micropeptides due to the
presence of short ORFs [40]. For example, Anderson et al. discovered a
highly conserved 46 amino-acid micropeptide, which the authors
denominated myoregulin, encoded by a skeletal muscle-specic RNA
annotated as a putative long non-coding RNA in human and mice.
Interestingly, this micropeptide was identied as an important
regulator of the Sarco-Endoplasmic Reticulum Ca2 +-ATPase (SERCA)
calcium pump in adult skeletal muscle [41].
Furthermore, strong evidence concerning circRNAs serving as templates for translation in vivo derived from Abe et al., who demonstrated
that an exonic circular RNA with an innite ORF is efciently translated
into living human cells to produce an abundant protein product by a
rolling-circle amplication mechanism [20] (Table 1). It is noteworthy
that translation of this circular RNA take place in the absence of any
particular element for internal ribosome entry, a poly-A tail, or a cap
structure in a eukaryotic system [20] (Fig. 2, right panel). This suggests
that there must be some elements within circularized eukaryotic
transcripts that can promote cap-independent translation initiation,
other than IRESs [42]. Furthermore, protein synthesis could occur
through a mechanism different from the canonical translation process,
such as initiation at non-AUG codons (e.g., initiation at CUG and GUG codons), leaky scanning, translational reinitiation, and/or translational
frameshifts [17,43], which also should be considered.
1249
these could only arise from circRNA translation; thus, the negative
ndings allowed to conclude that these brain expressed circRNAs, are
unlikely to be translated into peptides [55].
Practically all of the transcripts that have been suggested to encode
small peptides by ribosome proling [17] lack the evolutionary
conservation of their proposed coding regions [56,57], in contrast with
known protein-coding genes [58], including the few, well-characterized,
functional small peptides [53,59,60].
Accordingly, identication of any new protein-coding gene requires
additional biological information; therefore, further experimentation is
needed for demonstrating the existence and the role of the circRNAsderived protein in vivo [51,53].
4.2. Some bioinformatic tools for exploring circRNA coding potential
At present, to our knowledge, no computational resource is available
to systematically search for circRNAs with coding potential. The search
for ORFs into the circRNA sequence could provide insights into the function of circRNA-derived proteins. For this, the ORF Finder comprises a
graphical analysis tool that localizes all ORFs of a selectable minimal
size in a sequence provided by the user or in a sequence already located
in the PubMed database (Table 2).
Another approach could include homology search of a putative
product sequence within protein-domain databases (e.g., Pfam 29.0),
in an attempt to nd matches with reference proteomes sequences.
Thus, the identication of domains that occur within known proteins
can provide insights into their function [62] (Table 2). Additionally, bioinformatics tools, such as the PhyloCSF method [48], represent a useful
strategy for examining the coding potential of circRNAs through
evolutionary signatures characteristic of alignments of conserved
coding regions. This algorithm searches for matches with high
frequency of synonymous codon substitutions and conservative
amino-acid substitutions [47,48] (Table 2). Due to the ever-increasing
number of available genome sequences, these methods have been
successfully employed for accurate determination of conserved coding
potential in very small regions (of only ve amino acids) [56].
On the other hand, cellular IRES exist to play a crucial role at some
critical moments of cell life when cap-dependent translation initiation
is compromised [42]. Structures of many IRES are well-known and are
hypothesized to be important due to their function in translation
initiation [37]. Very recently, Dudekula et al. developed a new
computational resource named the CircInteractome that, in addition
to facilitate the search for possible interactions of human circRNAs
with RNA-binding proteins and miRNAs, allows the investigation of
potential circRNA translation through IRES sequences (Table 2). To
perform this, the tool compares all experimentally validated IRES
sequences [37] against mature circRNA sequences in order to identify instances of these cis-acting elements in circRNAs [63]. By means
of this resource, it was possible to nd that hsa_circ_0041407
contains an IRES and partial coding sequences of the MAX network
transcriptional repressor, and it was suggested that this circRNA
could give rise to a small chimeric protein of ~ 31 kDa [63]. With
this Web tool, a researcher is also assisted to design junctionspanning primers for the specic detection of circRNAs of
interest and to design small interfering RNAs for circRNA silencing.
Thus, with the data from current databases taken together, this
represents a great advantage for the analysis of the translational
role of circRNAs in vitro and in vivo (Table 2). However,
nucleotide composition and especially (G + C) content (introns
being more A/T-rich than exons, particularly in plants), codon
composition, hexamer frequency, base occurrence periodicity,
among others, are other measures that have been considered for
attempting the ne characterization of a sequence possessing coding
potential for translation into detectable peptides. Therefore, this
feature must be exploited by diverse genomic algorithms through
different methods [64].
1250
Table 2
Algorithms and tools for identifying circRNA-translation potential.
Resource
names
Web links
ORF Finder
IRESite
http://www.ncbi.nlm.nih.gov/gorf/gorf.html
http://iresite.org/
Remarks
Refs.
This tool identies all ORFs that a transcript possesses using the standard or alternative genetic codes
This database could be employed to evaluate which of the many IRESs published to datethat are
experimentally validatedcan be aligned with one or more mature circRNA sequences, supporting
their existence
CPAT
http://lilab.research.bcm.edu/cpat/index.php Calculates the likelihood with which a presumed ncRNA could encode for a peptide
Pfam 29.0
http://pfam.xfam.org/
This tool is useful for analyses of the sequence features of ORFs and also screens for the presence of
known protein motifs within an ORF
PhyloCSF
http://compbio.mit.edu/PhyloCSF
PhyloCSF is helpful for evaluating the coding potential of transcript models or individual exons in an
assembled genome that can be aligned with one or more informant genomes at appropriate
phylogenetic distances
CircInteractome http://circinteractome.nia.nih.gov/index.html This computational tool enables the prediction and mapping of binding sites for RBPs and
miRNAs on reported circRNAs
[37]
[61]
[62]
[48]
[63]
CAPT, Coding-Potential Assessment Tool; ORF, open reading frame; IRES, internal ribosomal entry site; ncRNA, non-coding RNA; RBPs, RNA-binding proteins; miRNA, microRNA; circRNA,
circular RNA.
1251
[41] D.M. Anderson, K.M. Anderson, C.L. Chang, C.A. Makarewich, B.R. Nelson, J.R. McAnally,
P. Kasaragod, J.M. Shelton, J. Liou, R. Bassel-Duby, E.N. Olson, A micropeptide encoded
by a putative long noncoding RNA regulates muscle performance, Cell 160 (2015)
595606.
[42] I.N. Shatsky, S.E. Dmitriev, I.M. Terenin, D.E. Andreev, Cap- and IRES-independent
scanning mechanism of translation initiation as an alternative to the concept of
cellular IRESs, Mol. Cell 30 (2010) 285293.
[43] J.F. Atkins, R.F. Gesteland, Recoding: Expansion of Decoding Rules Enriches Gene
Expression, Springer, New York, 2010.
[44] S. Petkovic, S. Muller, RNA circularization strategies in vivo and in vitro, Nucleic
Acids Res. 43 (2015) 24542465.
[45] M.E. Dinger, K.C. Pang, T.R. Mercer, J.S. Mattick, Differentiating protein-coding and
noncoding RNA: challenges and ambiguities, PLoS Comput. Biol. 4 (2008),
e1000176.
[46] N. Brockdorff, A. Ashworth, G.F. Kay, V.M. McCabe, D.P. Norris, P.J. Cooper, S. Swift, S.
Rastan, The product of the mouse Xist gene is a 15 kb inactive X-specic transcript
containing no conserved ORF and located in the nucleus, Cell 71 (1992) 515526.
[47] M.F. Lin, A.N. Deoras, M.D. Rasmussen, M. Kellis, Performance and scalability of
discriminative metrics for comparative gene identication in 12 Drosophila
genomes, PLoS Comput. Biol. 4 (2008), e1000067.
[48] M.F. Lin, I. Jungreis, M. Kellis, PhyloCSF: a comparative genomics method to
distinguish protein coding and non-coding regions, Bioinformatics 27 (2011)
i275i282.
[49] J.L. Tupy, A.M. Bailey, G. Dailey, M. Evans-Holm, C.W. Siebel, S. Misra, S.E. Celniker,
G.M. Rubin, Identication of putative noncoding polyadenylated transcripts in Drosophila melanogaster, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 54955500.
[50] T. Kondo, S. Plaza, J. Zanet, E. Benrabah, P. Valenti, Y. Hashimoto, S. Kobayashi, F.
Payre, Y. Kageyama, Small peptides switch the transcriptional activity of
shavenbaby during Drosophila embryogenesis, Science 329 (2010) 336339.
[51] M.I. Galindo, J.I. Pueyo, S. Fouix, S.A. Bishop, J.P. Couso, Peptides encoded by short
ORFs control development and dene a new eukaryotic gene family, PLoS Biol. 5
(2007), e106.
[52] N.T. Ingolia, S. Ghaemmaghami, J.R. Newman, J.S. Weissman, Genome-wide analysis
in vivo of translation with nucleotide resolution using ribosome proling, Science
324 (2009) 218223.
[53] M. Guttman, J.L. Rinn, Modular regulatory principles of large non-coding RNAs,
Nature 482 (2012) 339346.
[54] M. Guttman, P. Russell, N.T. Ingolia, J.S. Weissman, E.S. Lander, Ribosome proling
provides evidence that large noncoding RNAs do not encode proteins, Cell 154
(2013) 240251.
[55] X. You, I. Vlatkovic, A. Babic, T. Will, I. Epstein, G. Tushev, G. Akbalik, M. Wang, C.
Glock, C. Quedenau, X. Wang, J. Hou, H. Liu, W. Sun, S. Sambandan, T. Chen, E.M.
Schuman, W. Chen, Neural circular RNAs are derived from synaptic genes and regulated by development and plasticity, Nat. Neurosci. 18 (2015) 603610.
[56] M. Guttman, I. Amit, M. Garber, C. French, M.F. Lin, D. Feldser, M. Huarte, O. Zuk, B.W.
Carey, J.P. Cassady, M.N. Cabili, R. Jaenisch, T.S. Mikkelsen, T. Jacks, N. Hacohen, B.E.
Bernstein, M. Kellis, A. Regev, J.L. Rinn, E.S. Lander, Chromatin signature reveals over
a thousand highly conserved large non-coding RNAs in mammals, Nature 458
(2009) 223227.
[57] M. Guttman, M. Garber, J.Z. Levin, J. Donaghey, J. Robinson, X. Adiconis, L. Fan, M.J.
Koziol, A. Gnirke, C. Nusbaum, J.L. Rinn, E.S. Lander, A. Regev, Ab initio reconstruction of cell type-specic transcriptomes in mouse reveals the conserved multiexonic structure of lincRNAs, Nat. Biotechnol. 28 (2010) 503510.
[58] M. Clamp, B. Fry, M. Kamal, X. Xie, J. Cuff, M.F. Lin, M. Kellis, K. Lindblad-Toh, E.S.
Lander, Distinguishing protein-coding and noncoding genes in the human genome,
Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 1942819433.
[59] K. Hanada, X. Zhang, J.O. Borevitz, W.H. Li, S.H. Shiu, A large number of novel coding
small open reading frames in the intergenic regions of the Arabidopsis thaliana
genome are transcribed and/or under purifying selection, Genome Res. 17 (2007)
632640.
[60] J.P. Kastenmayer, L. Ni, A. Chu, L.E. Kitchen, W.C. Au, H. Yang, C.D. Carter, D. Wheeler,
R.W. Davis, J.D. Boeke, M.A. Snyder, M.A. Basrai, Functional genomics of genes with
small open reading frames (sORFs) in S. cerevisiae, Genome Res. 16 (2006) 365373.
[61] L. Wang, H.J. Park, S. Dasari, S. Wang, J.P. Kocher, W. Li, CPAT: coding-potential
assessment tool using an alignment-free logistic regression model, Nucleic Acids
Res. 41 (2013), e74.
[62] R.D. Finn, J. Mistry, J. Tate, P. Coggill, A. Heger, J.E. Pollington, O.L. Gavin, P.
Gunasekaran, G. Ceric, K. Forslund, L. Holm, E.L. Sonnhammer, S.R. Eddy, A.
Bateman, The Pfam protein families database, Nucleic Acids Res. 38 (2010)
D211D222.
[63] D.B. Dudekula, A.C. Panda, I. Grammatikakis, S. De, K. Abdelmohsen, M. Gorospe,
CircInteractome: a web tool for exploring circular RNAs and their interacting
proteins and microRNAs, RNA Biol. 13 (2016) 3442.
[64] C. Mathe, M.F. Sagot, T. Schiex, P. Rouze, Current methods of gene prediction, their
strengths and weaknesses, Nucleic Acids Res. 30 (2002) 41034117.
[65] G. Housman, I. Ulitsky, Methods for distinguishing between protein-coding and long
noncoding RNAs and the elusive biological purpose of translation of long noncoding
RNAs, Biochim. Biophys. Acta 1859 (2016) 3140.
[66] K. Kashi, L. Henderson, A. Bonetti, P. Carninci, Discovery and functional analysis of
lncRNAs: methodologies to investigate an uncharacterized transcriptome, Biochim.
Biophys. Acta 1859 (2016) 315.
[67] S.P. Barrett, J. Salzman, Circular RNAs: analysis, expression and potential functions,
Development 143 (2016) 18381847.