Post-Transcriptional Gene Regulation

Savva, Yiannis A, Laurent, Georges St, Reenan, Robert A, Yiannis A. Savva, Georges St. Laurent, Robert A. Reenan

Erik Dassi

Adenosine (A)-to-inosine (I) RNA editing is a fundamental posttranscriptional modification that ensures the deamination of A-to-I in double-stranded (ds) RNA molecules. Intriguingly, the A-to-I RNA editing system is particularly active in the nervous system of higher eukaryotes, altering a plethora of noncoding and coding sequences. Abnormal RNA editing is highly associated with many neurological phenotypes and neurodevelopmental disorders. However, the molecular mechanisms underlying RNA editing-mediated pathogenesis still remain enigmatic and have attracted increasing attention from researchers. Over the last decade, methods available to perform genome-wide transcriptome analysis, have evolved rapidly. Within the RNA editing field researchers have adopted next-generation sequencing technologies to identify RNA-editing sites within genomes and to elucidate the underlying process. However, technical challenges associated with editing site discovery have hindered efforts to uncover comprehensive editing site datasets, resulting in the general perception that the collections of annotated editing sites represent only a small minority of the total number of sites in a given organism, tissue, or cell type of interest. Additionally to doubts about sensitivity, existing RNA-editing site lists often contain high percentages of false positives, leading to uncertainty about their validity and usefulness in downstream studies. An accurate investigation of A-to-I editing requires properly validated datasets of editing sites with demonstrated and transparent levels of sensitivity and specificity. Here, we describe a high signal-to-noise method for RNA-editing site detection using single-molecule sequencing (SMS). With this method, authentic RNA-editing sites may be differentiated from artifacts. Machine learning approaches provide a procedure to improve upon and experimentally validate sequencing outcomes through use of computationally predicted, iterative feedback loops. Subsequent use of extensive Sanger sequencing validations can generate accurate editing site lists. This approach has broad application and accurate genome-wide editing analysis of various tissues from clinical specimens or various experimental organisms is now a possibility.