Preprint: epigenetic features predict which adenosines undergo A-to-I RNA editing

A preprint posted to bioRxiv on 27 July 2026 reports that local epigenetic features, principally the histone modification H3K36me3, are strong predictors of which adenosine residues in RNA are edited to inosine at high versus low frequency. Adenosine-to-inosine (A-to-I) editing is catalysed by ADAR (adenosine deaminase acting on RNA) enzymes and is one of the most common post-transcriptional RNA modifications in animals; the resulting inosine is read by the cell as guanosine, diversifying the transcriptome beyond what is encoded in the genome.

The authors integrated published multi-omics datasets from human cell lines and mouse embryonic tissues and trained machine learning classifiers to distinguish high-efficiency from low-efficiency editing sites based on local chromatin features rather than RNA sequence or ADAR expression alone. H3K36me3 — a histone mark associated with actively transcribed gene bodies — emerged as the most consistently predictive feature across species, tissues, and developmental stages examined.

The findings suggest a chromatin-level layer of regulation upstream of ADAR enzyme activity that may help explain why individual adenosines show markedly different editing rates even within the same transcript. The work is relevant to researchers studying RNA biology, epigenomics, and transcriptome diversification, and may have implications for understanding ADAR-related disease associations. The preprint has not yet undergone peer review.

Note: this is a preprint. The analysis and conclusions have not been independently peer reviewed and should be interpreted accordingly.