Comprehensive Mapping of Human dsRNAome Reveals Conservation, Neuronal Enrichment, and Intermolecular Interactions

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• The human transcriptome contains millions of A-to-I editing sites, often used to infer the presence of double-stranded RNAs (dsRNAs).
• Current methods for inferring dsRNA presence are limited by transcription levels, read depth, and ADAR expression, and cannot detect unedited dsRNAs.

Purpose of the Study:

• To develop a computational method, dsRNAscan, to overcome the limitations of inferring dsRNA presence from editing sites.
• To provide a comprehensive annotation of dsRNAs in the human genome.

Main Methods:

• Development and application of dsRNAscan, a novel computational tool for predicting dsRNAs directly from genomic sequences.
• Analysis of the genomic distribution and characteristics of predicted dsRNAs.

Main Results:

• dsRNAscan predicted approximately 5 million dsRNAs in the human genome.
• Widespread genomic distribution of dsRNAs associated with repetitive elements; sparse distribution of non-repetitive dsRNAs enriched at chromosomal tips.
• Identification of potentially immunogenic dsRNAs, including a few long, highly paired structures, with only one located in a 3'UTR.

Conclusions:

• dsRNAscan offers a robust method for identifying dsRNAs, independent of editing status or expression levels.
• The study provides the first comprehensive catalog of human dsRNA annotations, revealing insights into their genomic distribution and potential biological functions.
• The findings highlight the potential immunogenicity of certain dsRNA structures and their specific cellular localization.