RNA editing is a natural process that changes specific RNA letters after a gene has already been transcribed, helping cells fine-tune how genetic information is used. One of the most common forms of RNA editing in mammals is carried out by enzymes called ADARs, which convert adenosine bases into inosine. These edits can influence how RNAs are translated, how stable they are, and how the immune system responds to RNA molecules.

Researchers at the Paul-Ehrlich-Institute, Germany have developed a new computational tool called TSniffer to make it easier to detect and measure RNA editing events using RNA sequencing data. Many existing approaches rely on predefined editing sites or extensive filtering, which can miss unexpected or rare editing signals. TSniffer takes a different approach by scanning RNA-seq data with a rolling window strategy to identify editing sites in an unbiased way.

TSniffer concept and basic operation modes

(a) Screenshot of RNA-sequencing data shown in the integrated genome viewer (IGV). Alignment shows coverage plot (allele frequency threshold is set to 0.05) and aligned reads from imput BAM file. Below is the loaded output GFF file from TSniffer analysis. Detected TsRegions are shown as color-coded bars (gold: TC; teal: AG). (b) TSniffer first performs a site-wise analysis generating read count tables from the input BAM alignment summarizing target transition (Ts) and non-transition (nTs) counts for each position. (c) For each transition (Ts) type (AG, CT, GA, TC), TSniffer performs a sliding window analysis calculating significant occurrence of Ts types using Fisher’s exact test of query transitions (q-Ts) against query non-transitions (q-nTs), non-query transitions (nq-Ts), and non-query non-transitions (nq-nTs). For details see Methods section. (d) Overlapping significant windows are merged into TsRegions. RTF values, p-values, and TsSite counts are re-calculated for each window. (e) Summary information of a TsRegion. RTF_mean and p_mean are derived from the individual overlapping windows, whereas RTF_exact and p_exact are re-calculated for the final window. TsPos indicates the exact nucleotide positions of identified TsSites. (f) Simple workflow of TSniffer deNovo and TSniffer Regio analyses.

TSniffer can operate in two modes. In one mode, it identifies and quantifies RNA editing sites within individual samples. In the second mode, it measures editing activity across predefined regions in multiple datasets, making it useful for comparisons across conditions or experiments. This flexibility allows researchers to study both global editing patterns and targeted regions of interest.

To validate the method, the authors analyzed RNA sequencing datasets from normal cells and from cells lacking specific ADAR enzymes. By comparing these conditions, they confirmed that TSniffer accurately detects biologically meaningful RNA editing events and can distinguish edits linked to different ADAR proteins. The results show that the tool reliably captures both known and previously uncharacterized editing sites.

Overall, TSniffer provides researchers with a powerful and accessible way to explore RNA editing directly from RNA sequencing data. By enabling unbiased discovery and robust quantification of editing activity, the method helps deepen our understanding of how RNA editing shapes gene regulation in health and disease.

Availability – The TSniffer code is freely available for non-commercial use under GNU-AGPL v3.0 licence and can be downloaded from GitHub (https://github.com/maiher/tsniffer_1.0) or Zenodo (https://zenodo.org/records/17829712).