Figure 1.
Distribution and validation of editing sites observed in RNA-seq data.
A) Distribution of RNA editing types with and without a filter for significant strand bias from RNA-seq data. B,C) RNA-seq traces are shown for one canonical RNA editing site (C-to-U in Serinc1) and for one non-canonical editing site (G-to-C in Lars2). Reads sequenced in the sense direction are shown in dark grey, while reads sequenced in the reverse direction are shown in light grey. D,E) Sanger sequencing validation results for sites in Serinc1 and Lars2. F,G) RFLP validation of sites in Serinc1 and Lars2. Samples exposed to enzyme are labeled “en+” and control samples are labeled “en-.” H) Genomic distribution of editing sites and random background.
Figure 2.
Identification of high-confidence RNA editing sites.
A) Distribution of editing ratio for all observed canonical and non-canonical RNA editing sites in our adipose RNA-seq data shows that the ratio is higher for non-canonical than canonical sites. B) Distribution of Fisher's exact test p-values for strand bias. Non-canonical RNA editing sites shows an extreme peak around zero, indicating that most non-canonical RNA editing sites are supported by strand biased reads. C) Overview of our RNA editing analysis pipeline.
Figure 3.
Example of RNA editing at microRNA target sites.
One hundred nucleotides of the 3′UTR of Rpa1 is shown. Multiple microRNA target sites form a dense cluster in this region and contain many A-to-I editing sites. Red bases represent RNA editing sites, and blue and green bases represent different microRNA seed locations.