circRNADb: A comprehensive database for human circular RNAs with protein-coding annotations.
Chen X, Han P, Zhou T, Guo X, Song X, Li Y. · 2016
Read circRNADb: A comprehensive database for human circular RNAs with protein-coding annotations. by Chen X, Han P, Zhou T, Guo X, Song X, Li Y. (2016) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc5057092.txt.
Introduction
Xiaoping Chen, Ping Han, Tao Zhou, Xuejiang Guo, Xiaofeng Song, and Yan Li build a database of human circular RNAs in Scientific Reports in 2016. Circular RNAs are loop-shaped RNAs found across human tissues, with roles in health and disease, but no fully annotated collection existed. Their database, circRNADb, gathers 32,914 human exonic circular RNAs chosen carefully from many sources.
Each entry carries genome position, splicing pattern, sequence, and references. The striking feature is protein-coding annotation. 16,328 circles carry a reading frame longer than one hundred amino acids. 7,170 of them also bear an internal entry site that could start protein building, and mass checks found matching proteins for 46 circles from 37 genes. Circles were long classed as non-coding, so the coding clues open a new question. The site offers search, browsing, download, and a way to submit new finds, and the team plans steady updates.
Catalog record
| CREATOR | Chen X, Han P, Zhou T, Guo X, Song X, Li Y. |
|---|---|
| YEAR | 2016 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
circRNADb: A comprehensive database for human circular RNAs with protein-coding annotations By Xiaoping Chen, Ping Han, Tao Zhou, Xuejiang Guo, Xiaofeng Song, Yan Li Published in: Scientific Reports DOI: 10.1038/srep34985 ABSTRACT Abstract It has been known that circular RNAs are widely expressed in human tissues and cells, and play important regulatory roles in physiological or pathological processes. However, there is lack of comprehensively annotated human circular RNAs database. In this study we established a circRNA database, named as circRNADb, containing 32,914 human exonic circRNAs carefully selected from diversified sources. The detailed information of the circRNA, including genomic information, exon splicing, genome sequence, internal ribosome entry site (IRES), open reading frame (ORF) and references were provided in circRNADb. In addition, circRNAs were found to be able to encode proteins, which have not been reported in any species. 16328 circRNAs were annotated to have ORF longer than 100 amino acids, of which 7170 have IRES elements. 46 circRNAs from 37 genes were found to have their corresponding proteins expressed according mass spectrometry. The database provides the function of data search, browse, download, submit and feedback for the user to study particular circular RNA of interest and update the database continually. circRNADb will be built to be a biological information platform for circRNA molecules and related biological functions in the future. The database can be freely available through the web server at http://reprod.njmu.edu.cn/circrnadb. Unlike linear RNA, circular RNA is a special group of non-coding RNA which forms a covalently closed continuous loop from exon circularization. In classical molecular biology, precursor RNA produced from DNA template strand by transcription can be processed into mature linear messenger RNA by canonical RNA splicing, in which introns are removed, while exons connect together in genomic order. However, non-canonical splicing can make exons scrambled to form a circle1,2. The first circular RNA was recognized in the 1970s. In 1979, the researcher suggested that RNAs could exist in circular form in the cytoplasm of eukaryotic cells3. Ten years later, it was reported that human cytoplasmic RNA contained very low levels of transcripts of the DCC gene with scrambled exons4. For the next few decades, due to the specificity of the structure and the low expression level of circRNA, only a few genes were identified to express circRNAs, including DCC, EST-1, SRY etc. Recently, with the development of high throughput sequencing technology, a large number of circRNAs has been discovered across species5,6,7,8. These circRNA molecules were found to be evolutionary conservative, stable, and specifically expressed across tissues or developmental stages9,10,11,12. It has been shown that they play important roles in gene regulation9,13. Therefore circular RNA has become the hotspots in the current transcriptomics research field. Recently, as researchers put a lot of efforts into the study of circRNA, building a comprehensive circular RNA database become imperative. Several databases of the circRNA have been published, such as circBase, circRNABase and Circ2Trait14,15,16. The circBase merged and unified data sets of circRNAs from public references, with the evidence supporting their expression within the genomic context14. Circ2Traits is a comprehensive database for circRNA potentially associated with disease and traits15, which has only 1954 circRNAs. circRNABase is designed for decoding miRNA-circRNA interaction networks from thousands of circRNAs and 108 CLIP-Seq (HITS-CLIP, PAR-CLIP, iCLIP, CLASH) datasets16, however it does not provide the genomic information of circRNAs. In order to further study the circRNA and related biological functions, we build a comprehensive reference database, named as circRNADb. We collected dataset of circRNAs from relevant literatures, together with the circRNAs dataset identified from the Gliomas RNA-Seq dataset by our research group17. However, the primary data may have false positives (circRNAs with two ends from different genes) and redundancy, so we filtered the dataset according to gene annotation GTF file, and obtained a total of 32,914 human exonic circRNAs. Its detailed genomic information are also listed in the database, including its best matched transcript and the corresponding exon splicing information, genome sequences, in addition to all the possible isoforms and the corresponding exon splicing information. Although circRNA is classified as a non-coding RNA, researchers have reported that eukaryotic ribosome can initiate translation on circRNA, but only when the RNA contains internal ribosome entry site (or IRES) elements18. In 1995, Chen and colleagues showed that a synthetic circRNA containing IRES elements could recruit the ribosome to initiate translation, whereas those circRNAs without IRES did not18. Although the tested circRNA was a purely artificial construct, Chen and colleagues stated in their paper that they would be interested to see whether natural circRNAs contain IRES elements. So in this work we annotated the internal ribosome entry site and open reading frame (ORF) for the circRNA with protein-coding potential. Their protein expression evidences by mass spectrometry were also provided. Besides, we analyzed the features of the proteins translated from circRNAs, included domains, N-Glycosylation sites, mucin type O-Glycosylation sites and phosphorylation sites. Users can also employ “SProtP Human” to recognize those short-lived proteins based on sequence-derived features19. Finally, human circRNA data sets, along with its genomic features, protein coding potential and protein features were integrated into circRNADb. MATERIALS AND METHODS Raw circRNAs dataset in circRNADb were collected from related literatures9,10,11,12,17. We filtered those circRNAs supported by only one read, and only included those circRNAs supported by two or more reads. Basic genomic information about the circRNAs from the above dataset was extracted in BED format, including chromosome name, start position, end position and the cell (or tissue) type. Aided by gene annotation GTF file, we obtained the circRNA genomic information, such as strand, gene symbol, all possible transcription and the corresponding exon splicing information. The longest possible spliced transcript was taken as the best candidate sequence of the circRNAs. It has been known that canonical ribosome-based translation needs 5′-cap structure18. The alternative mechanism to initiate translation in eukaryotic mRNAs is that the RNA has internal ribosome entry site18. As circRNAs don’t have 5′-cap structure, another structure used for ribosome entry is internal ribosome entry site. An IRES element is a nucleotide sequence that allows the ribosome initiate translation directly in the middle of the mRNA sequence, instead of reading from the 5′ head to 3′ end20. If a circRNA contains at least one IRES element, it may be able to encode a protein. Thus, in order to annotate the protein-coding potential of all the circRNAs, we employed the method of VIPS depending on RNA structure similarity as proposed by Hong J. J. et al. to predict the IRES element in the spliced sequence of each circRNA21. VIPS has three key steps: RNA folding, RNA secondary structure comparison and pseudoknot prediction program21. RNALfold (from ViennaRNA package, version 2.1.9), RNA Align (provided by corresponding author of VIPS) and pknotsRG (version 1.3) were used in the IRES prediction system. RNALfold was used to predict local stable RNA secondary structures of long RNA by minimum free energy method22, while pknotsRG employed the newest Turner energy rules for finding the structure of minimal free energy, so it was able to predict a restricted class of pseudoknots23. In addition to IRES elements, the longest open reading frame was predicted for each circRNA. Any frame that starts with a start codon, and ends with a stop codon with a length longer