Accurate identification of RNA editing sites from primitive sequence with deep neural networks

Ouyang, Zhangyi; Liu, Feng; Zhao, Chenghui; Ren, Chao; An, Gaole; Mei, Chuan; Bo, Xiaochen; Shu, Wenjie

doi:10.1038/s41598-018-24298-y

Cited by 17 publications

(10 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the future, single cell-based transcriptome analysis can be used to identify nuances of differential ADAR editing landscapes, including between neural cell types within smaller brain regions and individual neural circuits. However, currently the accuracy of computational approaches of variant calling relies on a read depth greater than 15 (Ouyang et al, 2018) and most single cell datasets are currently performed using 10X genomics which is mostly below the threshold for accuracy. Since the majority of single cell RNA-seq experiments usually have read depths of no more than 10, they may not be as accurate in determining ADAR editing events.…”

Section: Discussionmentioning

confidence: 99%

Differential ADAR editing landscapes in major depressive disorder and suicide

Meindl

Piontkivska

2021

Preprint

View full text Add to dashboard Cite

Neuropsychiatric disorders, including depression and suicide, are becoming an increasing public health concern. Rising rates of both depression and suicide, exacerbated by the current COVID19 pandemic, have only hastened our need for objective and reliable diagnostic biomarkers. These can aide clinicians treating depressive disorders in both diagnosing and developing treatment plans. While differential gene expression analysis has highlighted the serotonin signaling cascade among other critical neurotransmitter pathways to underly the pathology of depression and suicide, the biological mechanisms remain elusive. Here we propose a novel approach to better understand molecular underpinnings of neuropsychiatric disorders by examining patterns of differential RNA editing by adenosine deaminases acting on RNA (ADARs). We take advantage of publicly available RNA-seq datasets to map ADAR editing landscapes in a global gene-centric view. We use a unique combination of Guttman scaling and random forest classification modeling to create, describe and compare ADAR editing profiles focusing on both spatial and biological sex differences. We use a subset of experimentally confirmed ADAR editing sites located in known protein coding regions, the excitome, to map ADAR editing profiles in Major Depressive Disorder (MDD) and suicide. Using Guttman scaling, we were able to describe significant changes in editing profiles across brain regions in males and females with respect to cause of death (COD) and MDD diagnosis. The spatial distribution of editing sites may provide insight into biological mechanisms under-pinning clinical symptoms associated with MDD and suicidal behavior. Additionally, we use random forest modeling including these differential profiles among other markers of global editing patterns in order to highlight potential biomarkers that offer insights into molecular changes underlying synaptic plasticity. Together, these models identify potential prognostic, diagnostic and therapeutic biomarkers for MDD diagnosis and/or suicide.

show abstract

Section: Discussionmentioning

confidence: 99%

Differential ADAR editing landscapes in major depressive disorder and suicide

Meindl

Piontkivska

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Detecting inosine formation in RNA is of central importance for characterizing editing mechanisms. While high‐throughput RNA sequencing (RNA‐seq) is commonly employed for large scale detection and mapping of A‐to‐I sites, this method is also costly, prone to random sampling errors, and requires complex bioinformatic analyses . Alternatively, model reactions using ADAR enzymes with small RNA substrates (≈20—50 nt) have yielded substantial insights into how certain RNA sequences and structural motifs are recognized and edited .…”

Section: Methodsmentioning

confidence: 99%

Chemical Profiling of A‐to‐I RNA Editing Using a Click‐Compatible Phenylacrylamide

Knutson

Korn

Johnson

et al. 2020

Chemistry A European J

View full text Add to dashboard Cite

Straightforwardm ethods for detecting adenosine-to-inosine (A-to-I) RNA editinga re key to ab etter understanding of its regulation, function, and connection with disease. Wea ddress this need by developing an ovel reagent, N-(4-ethynylphenyl)acrylamide (EPhAA), and illustrating its ability to selectively label inosine in RNA. EPhAA is synthesized in as ingle step, reacts rapidlyw ith inosine, and is "click"-compatible, enabling flexible attachment of fluorescent probesa te diting sites. We first validate EPhAA reactivity and selectivityf or inosine in both ribonucleosides and RNA substrates, and then apply our approach to directly monitor in vitro A-to-I RNA editing activity using recombinantA DAR enzymes. This method improves upon existing inosine chemical-labeling techniques and provides ac ost-effective, rapid, and non-radioactive approach for detecting inosine formation in RNA. We envision this method will improve the study of A-to-I editing and enable better characterization of RNA modification patterns in different settings. RNA is chemically modified by anumber of enzymes after transcription, in turn influencing RNA stability, localizationa nd activity within the cell. Adenosine-to-inosine (A-to-I) RNA editing is one of the most widespread modifications, and is performed by adenosine deaminases acting on RNA (ADARs) (Scheme 1a). [1] Adenosine deaminationc hanges the molecular structurea nd hydrogen-bonding pattern of the nucleobase, and resulting inosines insteadb ase pairw ith cytidinet oe ffectively recode these sites as guanosine. Editings ites within protein-coding mRNAsd irectly alter amino acid sequences and produce different protein isoforms. Non-coding RNAs also undergo extensive editing, including microRNAs and small-interfering RNAs, significantly alteringt heir biosynthesis, localization, and gene regulation properties. [2-3] A-to-I editing is essential for an umber of biological processes including tissue development, [4-5] neurologicalf unction, [6] and immune system activation. [7] Dysfunctional editing is also directlyl inked with autoimmune diseases, [8-9] neurological disorders, [10] and several types of cancer. [11-12] Despite this importance,o ur overall understanding of A-to-I editingr egulation is limited. In particular,w hile many sites have been identified (> 5million), [13-14] it is unclear why certain sites are edited at higherf requency than others and what precise function they each serve. [15] Efforts to map A-to-I locations and ADAR binding sites have revealed that editingp atterns are highly complex and variable in humans, [7, 16-18] and the precise mechanismsb yw hich ADAR enzymes bind to and edit specific RNA sequences remain unclear.T his gap is also significant for therapeutic site-directed RNA editing strategies, [19] as both the design and precise implementation of this machinery is reliant on at horough understanding of ADAR regulation. Detecting inosine formation in RNA is of central importance for characterizing editingm echanisms. While high-throughput RNA sequencin...

show abstract

“…the model, or overall nucleotide content and codon information, can explain 23% of the variation in the extent of editing), which is not high but is expected since extent of editing depends not only on overall properties of the gene, but also specific sequence information. In fact, previous studies have been able to predict specific editing sites from SNVs with very high accuracy, but these models rely on the specific sequence surrounding the site and have only been tested on data from human and Drosophila ( 31 ). The predictive ability of our models, given that they only consider gene-level properties, thus indicate that indeed overall nucleotide content and codon usages are related to RNA editing.…”

Section: Predicting Editing With Nucleotide Composition and Codon Usagesmentioning

confidence: 97%

Evolutionary Analyses of RNA Editing and Amino Acid Recoding in Cephalopods

Wang

Mohlhenrich

2019

Preprint

View full text Add to dashboard Cite

RNA editing is a post-transcriptional modification process that alters nucleotides of mRNA and consequently the amino acids of the translated protein without changing the original DNA sequence. In human and other mammals, amino acid recoding from RNA editing is rare, and most edits are non-adaptive and provide no fitness advantage ( 1 ). However, recently it was discovered that in soft-bodied cephalopods, which are exceptionally intelligent and include squid, octopus, and cuttlefish, RNA editing is widespread and positively selected ( 2 ). To examine the effects of RNA editing on individual genes, we developed a "diversity score" system that quantitatively assesses the amount of diversity generated in each gene, incorporating combinatorial diversity and the radicalness of amino acid changes. Using this metric, we compiled a list of top 100 genes across the cephalopod species that are most diversified by RNA editing. This list of candidate genes provides directions for future research into the specific functional impact of RNA editing in terms of protein structure and cellular function on individual proteins. Additionally, considering the connection of RNA editing to the nervous system, and the exceptional intelligence of cephalopod, the candidate genes may shed light to the molecular development of behavioral complexity and intelligence. To further investigate global influences of RNA editing on the transcriptome, we investigated changes in nucleotide composition and codon usage biases in edited genes and coleoid transcriptome in general. Results show that these features indeed correlate with editing and may correspond to causes or effects of RNA editing. In addition, we characterized the unusual RNA editing in cephalopods by analyzing ratio of radical to conservative amino acid substitutions (R/C) and distribution of amino acid recoding from editing. Our results show that compared to model organisms, editing in cephalopods have significantly decreased R/C ratio and distinct distribution of amino acid substitutions that favor conservative over radical changes, indicating selection at the amino acid level and providing a potential mechanism for the evolution of widespread RNA editing in cephalopods.

show abstract

Accurate identification of RNA editing sites from primitive sequence with deep neural networks

Cited by 17 publications

References 54 publications

Differential ADAR editing landscapes in major depressive disorder and suicide

Differential ADAR editing landscapes in major depressive disorder and suicide

Chemical Profiling of A‐to‐I RNA Editing Using a Click‐Compatible Phenylacrylamide

Evolutionary Analyses of RNA Editing and Amino Acid Recoding in Cephalopods

Contact Info

Product

Resources

About