Fig. 3. Tissue-dependence and ADAR-specificity of CDS editing sites.

a Number of sites detected and edited to >1% per-tissue (out of the total 1517). The highest numbers of sites are observed in the nervous system, arteries, and lung. b Box-and-whisker plots, depicting the distribution of per-tissue CDS editing index values reveal that although the number of edited sites in the brain is large, CDS editing activity (number of deamination events) is comparable to most other tissues and is much lower than in arteries, colon or esophagus. Box-and-whisker plots show the medians (horizontal lines), upper and lower quartiles (box edges), and 1.5 × the interquartile range (whiskers). See Supplementary Data 1 for the number of biologically independent samples per tissue. c t-SNE¹¹⁸ dimensionality-reduction analysis (“Methods”) reveals highly distinctive CDS editing pattern in cerebellum, arteries, testis, and non-cerebral brain regions. d The number of sites detected depends strongly on the editing level cutoff. For all tissues, about half of the detected sites are weakly edited (<1%). e Analyzing previously published RNA-seq data from ADAR1- and ADAR2-overexpressing human cell lines, we classify 227 of the sites (well covered and sufficiently edited in cell lines data) based on the editing enzyme. The vast majority of these are targeted by ADAR2 (see Supplementary Data 8). (f) Analyzing previously published RNA-seq data from ADAR1 and ADAR2 knockout mice^14,49, we calculate the ADAR Influence value (see Methods) for 58 sites conserved in mouse (all sites with >5% in wild type mouse and significant, Fisher p < 0.05, difference between wild type and the double-knockout mouse, excluding the GRIA2 site that was genomically edited to G in the double-knockout) (See Supplementary Data 9). The heatmaps depict the editing levels and the ADAR Influence value (darker: ADAR1, lighter: ADAR2). g The biological variability (“Methods”) in editing levels across individual donors. Many of the strongly edited sites exhibit standard deviations much smaller than the mean, implying regulation on the RNA editing activity. Source data are provided as a Source Data file.