Summary

A bioRxiv preprint introduces RADIS, a chemical-probing strategy that maps RNA secondary structure across long, intron-rich genomic loci. The study applies it to splice sites, branchpoints and Alu elements, including deep intronic regions that are difficult to analyse experimentally.

Researchers have introduced RADIS, a method designed to measure RNA secondary structure across long, intron-rich genomic regions rather than only around known splice sites. In a bioRxiv preprint posted on September 13, 2026, the authors report that the approach produces strand-resolved structural data across regions containing splice sites, branchpoints, disease-associated intronic sequences and Alu elements.

Contents

Why deep intronic structure matters

Pre-messenger RNA, or pre-mRNA, is the initial RNA copy made from a gene. It contains exons, which are retained in the mature messenger RNA, and introns, which are generally removed during splicing. Before and during this process, the RNA strand can fold back on itself as bases pair to form secondary structures.

Those structures can affect how accessible important sequence features are. A folded region may mask a splice-site signal, while a more open region may expose it to the spliceosome, the molecular machinery that removes introns. Introns also contain branchpoints, sequence regions used during the chemical steps of intron removal.

The authors note that experimentally supported structural models for human pre-mRNAs have largely been concentrated near splice sites. Deep intronic regions have therefore been much less comprehensively mapped, even though the preprint estimates that roughly 70% of intronic disease-causing variants occur in these deeper regions.

How RADIS measures RNA structure

RADIS stands for Reactivity Analysis of Deep and Intergenic RNA Structure. The strategy creates tiled arrays of RNA molecules that collectively cover long genomic distances. These overlapping RNA segments allow chemical probing to be extended across intron-rich loci, where conventional approaches can be limited by RNA abundance and by the difficulty of targeting individual regions.

Chemical probing records how readily different nucleotides react with the probing chemicals. In general, nucleotides that are exposed or less constrained by base pairing show different reactivities from nucleotides involved in folded structures. The resulting profiles can then be used to constrain computational models of RNA folding. Because RADIS produces strand-resolved profiles, it can distinguish structures formed from the two possible directions of transcription.

The researchers tested the approach against known E. coli ribosomal RNA architectures. They also compared RADIS measurements with in-cell dimethyl sulfate mapping, or DMS-MaP, at representative splice sites. The reported correlation was high, with correlation coefficients ranging from 0.87 to 0.90.

What the analysis found

Using RADIS-constrained folding, the researchers examined 109 5′ splice sites and 88 3′ splice sites. They found an inverse relationship between base pairing within the region occupied by the spliceosome and splice-site strength: stronger splice sites tended to have less base pairing in that footprint.

The analysis also divided 81 branchpoints into three structural classes. This provides a way to compare how the local RNA environment around these sites differs, rather than treating branchpoint sequences as purely linear strings of bases.

A separate analysis covered 233 full-length Alu elements. Alu elements are repeated genomic sequences that can be transcribed in either the same orientation as the surrounding locus or the opposite orientation. The sense-oriented Alu RNAs were more structured than the antisense elements in the authors’ analysis. Both orientations could be divided into structural classes associated with strand and evolutionary lineage.

A resource for studying variant effects

The main contribution of RADIS is an experimental framework for examining RNA structure in regions that are often difficult to study directly. That matters for intronic and intergenic loci because a DNA variant can potentially affect not only the sequence of a regulatory signal, but also the way the surrounding RNA folds and presents that signal to cellular machinery.

The authors position RADIS as a complement to sequence-based variant-effect predictors. Measured reactivity profiles could provide structural information for models assessing disease-associated variants, particularly in deep intronic regions. The preprint presents this as a foundation for future studies of how specific sequence changes alter RNA structure and splicing.

The work is currently a bioRxiv preprint. Its reported contribution is a large-scale structural mapping method and the patterns identified from applying it to splice sites, branchpoints and Alu elements; connecting individual structural changes to patient outcomes or clinical applications will require further study.

Sources