RED-ML: a novel, effective RNA editing detection method based on machine learning

Xiong, Heng; Liu, Dongbing; Li, Qiye; Lei, Mengyue; Xu, Liqin; Wu, Liang; Wang, Zongji; Ren, Shancheng; Li, Wangsheng; Xia, Min; Lü, Lihua; Lu, Haorong; Hou, Yong; Zhu, Shida; Liu, Xin; Sun, Ying; Wang, Jian; Yang, Huanming; Wu, Kui; Xu, Xun; Lee, Leo J.

doi:10.1093/gigascience/gix012

Cited by 37 publications

(32 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Here, we used hisat 2, as the efficiency of this aligner has already been highlighted in aligning reads to the genome for predicting RNA editing sites (Xiong et al . ; Yao et al . ).…”

Section: Discussionmentioning

confidence: 99%

Large‐scale potential RNA editing profiling in different adult chicken tissues

2019

View full text Add to dashboard Cite

RNA editing is a post-transcription maturation process that diversifies genomically encoded information and can lead to transcriptome diversity. Thanks to next-generation sequencing technologies, a large number of editing sites have been identified in different species. Although this mechanism is well described in mammals, only a few studies have been performed in chicken. Here, candidate or potential RNA editing sites were identified in eight different tissues of chicken (brain, spleen, colon, lung, kidney, heart, testes and liver). We identified 68 A-to-G editing sites in 46 genes. Only two of these were previously reported in chicken. We found no C-to-T sites, attesting to the lack of this type of editing mechanism in chicken. Similar to mammals, the editing sites were enriched in non-coding regions, rarely resulted in a change in amino acids, showed a critical role in the nervous system and had a low guanosine level upstream of the editing site and some enrichment downstream from the site. Moreover, in contrast to mammals, editing sites were weakly enriched in interspersed repeats and the number and editing ratio of non-synonymous sites were higher than for those of synonymous sites. Interestingly, we found several tissue-specific edited genes, including GABRA3, SORL1 and HTR1D in brain and RYR2 and FHOD3 in heart, that were associated with functional processes relevant to the corresponding tissue. This finding highlights the importance of RNA editing in several chicken tissues, especially the brain, and establishes a foundation for further exploration of this process.

show abstract

“…Here, we used hisat 2, as the efficiency of this aligner has already been highlighted in aligning reads to the genome for predicting RNA editing sites (Xiong et al . ; Yao et al . ).…”

Section: Discussionmentioning

confidence: 99%

Large‐scale potential RNA editing profiling in different adult chicken tissues

2019

View full text Add to dashboard Cite

show abstract

“…A list of exon-exon junctions extracted from the known gene model annotation (Ensembl release 84) was used to guide the read mapping. Notable features of the Hisat2 program are that, 1) it prevents reads from being aligned to pseudogenes, which results in improved alignment accuracy [ 32 ] and 2) it is more efficient at providing editing prediction from RNA-Seq data than other programs [ 33 ]. We considered only uniquely and concordantly paired-end mapped reads, to reduce the potential bias caused by short read alignment.…”

Section: Methodsmentioning

confidence: 99%

Genome-wide identification and analysis of A-to-I RNA editing events in bovine by transcriptome sequencing

2018

View full text Add to dashboard Cite

RNA editing increases the diversity of the transcriptome and proteome. Adenosine-to-inosine (A-to-I) editing is the predominant type of RNA editing in mammals and it is catalyzed by the adenosine deaminases acting on RNA (ADARs) family. Here, we used a largescale computational analysis of transcriptomic data from brain, heart, colon, lung, spleen, kidney, testes, skeletal muscle and liver, from three adult animals in order to identify RNA editing sites in bovine. We developed a computational pipeline and used a rigorous strategy to identify novel editing sites from RNA-Seq data in the absence of corresponding DNA sequence information. Our methods take into account sequencing errors, mapping bias, as well as biological replication to reduce the probability of obtaining a false-positive result. We conducted a detailed characterization of sequence and structural features related to novel candidate sites and found 1,600 novel canonical A-to-I editing sites in the nine bovine tissues analyzed. Results show that these sites 1) occur frequently in clusters and short interspersed nuclear elements (SINE) repeats, 2) have a preference for guanines depletion/enrichment in the flanking 5′/3′ nucleotide, 3) occur less often in coding sequences than other regions of the genome, and 4) have low evolutionary conservation. Further, we found that a positive correlation exists between expression of ADAR family members and tissue-specific RNA editing. Most of the genes with predicted A-to-I editing in each tissue were significantly enriched in biological terms relevant to the function of the corresponding tissue. Lastly, the results highlight the importance of the RNA editome in nervous system regulation. The present study extends the list of RNA editing sites in bovine and provides pipelines that may be used to investigate the editome in other organisms.

show abstract

“…We collected 132 rhythmic editing sites, mediated by the rhythmic expression of ADARB1 in mouse liver ( 36 ). In addition, we analyzed RNA editing sites in seven circadian transcriptome datasets from 14 mouse tissues and two human tissues using RED-ML ( 53 ). Rhythmic editing sites were identified using LSPR ( 47 ) and annotated using HOMER ( 51 ).…”

Section: Data Collection and Processingmentioning

confidence: 99%

CirGRDB: a database for the genome-wide deciphering circadian genes and regulators

Shi

Zhang

et al. 2017

Nucleic Acids Research

View full text Add to dashboard Cite

Circadian rhythms govern various kinds of physiological and behavioral functions of the living organisms, and disruptions of the rhythms are highly detrimental to health. Although several databases have been built for circadian genes, a resource for comprehensive post-transcriptional regulatory information of circadian RNAs and expression patterns of disease-related circadian RNAs is still lacking. Here, we developed CirGRDB (http://cirgrdb.biols.ac.cn) by integrating more than 4936 genome-wide assays, with the aim of fulfilling the growing need to understand the rhythms of life. CirGRDB presents a friendly web interface that allows users to search and browse temporal expression patterns of interested genes in 37 human/mouse tissues or cell lines, and three clinical disorders including sleep disorder, aging and tumor. More importantly, eight kinds of potential transcriptional and post-transcriptional regulators involved in the rhythmic expression of the specific genes, including transcription factors, histone modifications, chromatin accessibility, enhancer RNAs, miRNAs, RNA-binding proteins, RNA editing and RNA methylation, can also be retrieved. Furthermore, a regulatory network could be generated based on the regulatory information. In summary, CirGRDB offers a useful repository for exploring disease-related circadian RNAs, and deciphering the transcriptional and post-transcriptional regulation of circadian rhythms.

show abstract

RED-ML: a novel, effective RNA editing detection method based on machine learning

Cited by 37 publications

References 33 publications

Large‐scale potential RNA editing profiling in different adult chicken tissues

Large‐scale potential RNA editing profiling in different adult chicken tissues

Genome-wide identification and analysis of A-to-I RNA editing events in bovine by transcriptome sequencing

CirGRDB: a database for the genome-wide deciphering circadian genes and regulators

Contact Info

Product

Resources

About