CircNet Tutorial

OVERVIEW

MAIN IINTERFACE

ADVANCED FUNCTIONS

BLAST

EXAMPLE APPLICATION


OVERVIEW


Circular RNAs (circRNAs) represent a new type of regulatory noncoding RNA that only recently has been identified and cataloged. Emerging evidence indicates that circRNAs exert a new layer of post-transcriptional regulation of gene expression. In this study, we utilized transcriptome sequencing datasets to systematically identify the expression of circRNAs (including known and newly identified ones by our pipeline) in 464 RNA-seq samples, and then constructed the CircNet database (http://circnet.mbc.nctu.edu.tw/) that provides the following resources: (i) novel circRNAs, (ii) integrated miRNA-target networks, (iii) expression profiles of circRNA isoforms, (iv) genomic annotations of circRNA isoforms (e.g., 282,948 exon positions), and (v) sequences of circRNA isoforms. The CircNet database is to our knowledge the first public database that provides tissue-specific circRNA expression profiles and circRNA-miRNA-gene regulatory networks. It not only extends the most up to date catalog of circRNAs but also provides a thorough expression analysis of both previously reported and novel circRNAs. Furthermore, it generates an integrated regulatory network that illustrates the regulation between circRNAs, miRNAs and genes.


Key Features


  1. Provides a total of 212,950 circRNAs including 53,687 novel RNAs. To reduce the false positives of circRNAs, we defined high-confidence circRNAs by selecting circular junction sites with supporting experiments >= 3. It resulted in 34,000 high-confidence circRNAs.

  1. Collects 464 RNA-seq samples (without PolyA selection) covering 26 different human tissues, and provides an interface to show circRNA expression profiles across these human tissues.

  1. Predicts circRNA-miRNA interactions and integrates with miRNA-target interactions to generate circRNA-miRNA-gene regulatory networks.

  1. Provides genomic annotation of circRNAs with a genome browser integrated into the web interface.



Database Construction


Framework of the database construction in CircNet.

The graph illustrates how the network in CircNet was constructed. For circRNA identification, transcriptome sequencing data sets were obtained from the NCBI Sequence Read Archive (SRA). The back splice junction sites in each RNA-Seq sample were identified using a circRNA discovery pipeline we developed (scripts provided on circBase). Detected back spliced junction sites were further compared with hg19 human genome annotation to define circRNA isoforms. Only isoforms with an average expression level, represented as average FPKM in 464 examined samples, greater than 0.01 were included in the database. To predict circRNA-miRNA interactions, the occurrence of miRNA target seeds in circRNA isoforms were examined and normalized by isoform length. The significance of interactions were evaluated by referring to the background distribution of miRNA seeds in all transcripts and only circRNA-miRNA interactions with P < 0.001 were reported.



Isoforms Identification


Identification of circRNA isoforms

 (A) As back-spliced junction site flanking exons may be composed of multiple isoforms, we included all the possible isoforms into the annotation for the expression analysis and miRNA target search. (B) For those circRNAs found with small differences to exon locations, flanked exons and a small portion of intron sequence were considered as parts of the isoforms. (C) Some of the back-spliced junction sites, despite overlapping with certain genes, localize to their antisense strands. For these circRNAs, we took the entire antisense flanked sequence between the “head” and “tail” position. (D) If the back-spliced junction sites were located within intergenic regions, the entire strip of flanked sequence was considered as part of the circRNAs.



microRNA Sponge Site Identification


MicroRNA Sponge Detection

Potential miRNA binding sites on circRNAs were identified through iteratively searching the circRNA isoform sequences for miRNA target sequences deemed typical: 6mer, 7mer-A1, 7mer-m8 and 8mer sequences(Bartel). Perfect complementarity was required of these sequences for a circRNA-miRNA target relationship to be identified. To normalize the number of occurrences of these sites, the below formula was used:

With this formula, four frequency numbers can be acquired from each pair of circRNA and miRNA. To distinguish circRNA from linear isoforms, frequency values were also calculated for linear mRNA and miRNA pairs. P-values for each circRNA and miRNA relationship were acquired by one minus the calculated the cumulated distribution function of the frequency Z score. The circRNA-miRNA pair with P-value < 0.001 represents high regulatory potential between the circRNA and miRNA. Information of these regulatory relationships was combined with target genes of miRNA from miRTarBase to form the networks demonstrated on CircNet.



Benchmarks


Comparison of CircNet, starBase and circBase

Feature

circBase

starBase v2.0

CircNet

Sequence alignment

A web interface of BLAT

Not available

A web interface of BLAST

Reference source of circRNAs

Back spliced junction sites in animals reported in 2013 and 2015.

circBase V0.1

Reported human circRNAs from 2013 to 2015.

Position of circRNAs on genome

Linked out to UCSC Genome Browser.

A customized genome browser accessible through keyword search.

An integrated genome browser synchronizing with the network graphical user interface.

Sample source of the circRNAs

The samples where the back-spliced junction sites were discovered.

circBase V0.1 source samples

1.    In which sample junction sites were discovered.

2.    Expression level in available samples.

3.    Clustered sample conditions.

Naming of circRNAs

A serial number for every detected back-spliced junction site.

Same as circBase, except CDR1 antisense (CDR1as).

A systematic naming system which provides:

1.    Information to the source gene.

2. Whether the circRNA is antisense or intronic.

3.    Well annotated exons of circRNAs.

CircRNA expression profiles in samples

Not available

Not available

An all-sample expression heat-map for every circRNA and linear isoform.

Address on miRNA regulatory relationships

Not available

Identifies circRNA and miRNA interactions through Chip-Seq data analysis.

A network-driven graphical interface shows the relationship between miRNA target genes and circRNAs.

Flanking repeats of circRNAs

Not available

Not available

Coordinates are traceable for all circRNAs and their linear isoforms for the human hg19 genome.

Isoforms of circRNAs

Not available

Not available

All are traceable on the integrated genome browser.



Development Team

Research and development laboratories

CirNet was developed through the joint effort of :



MAIN INTERFACE


Users can search gene or miRNA of interest, in the default setting:

Homepage of CircNet is composited of  three panels as follows:


A. Left panel: circRNA-miRNA-gene Regulatory Network

In this panel, post-transcriptional regulatory relationship of

was summarized in the network.  Double negative feedback control loops of circRNAs to genes can be found in this network.

This network would synchronize with user inputs.

User can search through:

Any interested circRNAs in this panel is clickable.

Whenever user clicks on one of the circRNAs, edges and nodes directly linked to the circRNAs would be highlighted, meanwhile panel B and panel C would synchronize with this action and demonstrate related detail information of the selected circRNA.

Legends is set as illustrated in the figure. In this network,

for the nodes :

for the edges:

User can set the threshold used to limit the amount of putative circRNAs through the Support drop down menu on the top left corner.


B. Top-right panel: circRNA profiles

Available detail information about selected circRNA is provided in this panel, three categories:


C. Bottom-right panel: Genome browser for circRNAs


In this panel, a genome browser synchronizing with user search is integrated.The genome browser mainly contains three tracks:

The scope of the genome browser would zoom to the location of the selected circRNA on the genome. The user is also free to select interested region on the browser. Sequences of the tracks can be acquired through clicking on them.


Please click on the refresh button on the top left corner whenever facing difficulty acquiring the network. The network would take a slightly longer time to show up for the first time use.



Whole Screen View

All three panels can be expanded to a full screen views through clicking on the top right buttons.


Export Graphs


All networks in CircNet can be exported into png files through clicking on the ”Export” button:



ADVANCED FUNCTIONS

CirNet provides a series of advanced functions to suit the user’s interests on circRNA research, including:


Setting the threshold of circRNAs demonstration

User can set the threshold used to limit the amount of putative circRNAs through the Support drop down menu on the top left corner. Through these options, user can select to see only those circRNAs with up to more or equal to 10 experiments supporting evidences.Choosing higher threshold would reduce the amount of circRNAs showed on the screen but increase the credibility of the analysis. The default setting of CircNet only shows those circRNAs with multiple back-splicing junction sites spanning reads reported or found in at least three different experiments:



Checking the detail information of the selected circRNA


When users click a circRNA in the left panel (A), CircNet will show a circRNA-miRNA-gene regulatory network centered by the given circRNA in the top-right panel (B). Since only those miRNAs targeting the searched gene would show up on the screen, double negative feedback control loops of circRNAs to genes can be found in this network. To access to further details about particular circRNA, simply click on the node representing the interested circRNA. Upon clicking on the node, edges and nodes directly associated with the selected node would be highlighted:

Following this action, the three categories of information about the selected circRNA would show up in the panel (B). Meanwhile, scope of the genome browser would be zoomed into the exact location of the circRNA on human genome.  

Information in the top-right panel (B) available is detailed as follows:

  1. Expression view: expression profiles of circRNAs for 26 types of human tissues
  2. Sponge view: an extend network focus on selected circRNA
  3. Site source table: a table showing in which experiment that multiple back splice junction sites reads were found.



Expression view: expression profiles of circRNAs


The top right panel, after clicking the magnifier icon, would expand to a full screen view:

To zoom into certain interested type of tissue, for example, brain, on the expression profile, simply click on the number represents the tissue listed in the bottom of the penal. After clicking on the column number, the screen would switch into the expression profile with further details, focusing solely on the selected tissue originated samples. In this example, only the 31 samples originated from human brain tissue were demonstrated on the clustered heatmap:

The panel shows all available  samples of brain, click any item below the framework,it will link SRA in the NCBI.  To going back to the previous view, simply clicking on the “Back” bottom on the top left corner. Values of FPKM would show up whenever user moving the mouse cursor on the heatmap data points.



Sponge view: an extend network focus on selected circRNA


In this panel, users can access to all the putatively circRNA sponged miRNAs. Genes targeted by there miRNAs can also be accessed here:  

Legends of this network is the same as panel (A)In this network:

for the nodes :

for the edges:

This network, like all other network graphs available in CircNet, can be exported into png files through clicking on the ”Export” button

Amount of the putatively sponged miRNAs to be showed on the network can be selected through the drop down menu. The descending order of this option is based on the statically analysed chance based on amount of perfectly aligned miRNA binding sites found on the circRNA sequence.    


Site source table: multiple back splice junction sites reads


In this table, information about the back spliced junction site spanning the selected circRNA is provided. The information includes:



Back Spliced Junction Depth


Starting from 12/16 2015, the user can check the values of normalized counts of reads spanning the back spliced junction sites. Information available within this panel synchronize with user search inputs just like other panels of CircNet. Depth of the back splice junction sites, aka head and tail position of the circRNAs of the specifically searched gene would be demonstrated in this panel. The value SRPBM stands for  spliced reads per billion mapping as mentioned in the work of Jeck et al. 2013. SRPBM represents the count of reads spanning the site normalized by read length and number of reads mapped on human genome for each run. The detected value within non poly A selected RNA-Seq samples and ploy A selected samples were separated into two different columns. As shown in the figure, back spliced junction  sites count is significantly depleted within poly A selected samples, which reflects the fact that circRNAs are lack of ploy-A tails.  



BLAST


Sequence alignment search is also available on CircNet. Through integrated BLAST, users can test whether their interested nucleotide sequences can align to circRNAs or not. Back splice junction sites overlapped with only intergenic regions can possibly be accessed through this function.  

In BLAST function, users can input a DNA or RNA sequence to search circRNA isoforms included in CircNet.

The figure is the result of blast.



EXAMPLE APPLICATION


We use the ZEB1-derived network to illustrate the information provided by CircNet.

When ZEB1 is inputted as the keyword, CircNet collects available information about miRNAs targeting ZEB1, circRNAs originating from the ZEB1 sequence, and the regulatory relationship between the circRNAs and miRNAs to create in an integrated mRNA-miRNA-circRNA regulatory network. The network suggests that the circRNAs circ-ZEB1.5, circ-ZEB1.19, circ-ZEB1.17 and circ-ZEB1.33 have the potential to sponge the miRNA miR-200a-3p. Since miR200a has been reported to target ZEB1, this suggests the existence of a negative feedback control loop. The integrated genome browser provides the visualization of these circRNAs: all these circRNAs contain the exon in position chr10:31791276-31791437, which is enriched with miR-200a targeting seeds. From the flanking area of this specific exon, we found the flanking MIR repeat sequences MIR_dup1543 and MIR3_dup768 and MIR3_dup769, providing a lead to the biogenesis of the circRNA. As illustrated in the this figure:

expression patterns provided in CircNet indicate that all four identified circRNAs— circ-ZEB1.5, circ-ZEB1.19, circ-ZEB1.17 and circ-ZEB1.33—were up-regulated in available normal lung tissue samples compared to the lung cancer samples.