Extended Data Figure 3. Analysis of adult mouse tissues by mmPCR–seq.

a, Average editing levels of sites at coding and untranslated region (UTR) positions in 12 mouse tissues from a single individual (129S1 strain). b, Correlations between ADAR expression levels (quantified as the number of RNA-seq fragments per kilobase of transcript per million mapped reads (FPKM)) and overall editing levels in different mouse tissues. The overall editing level is defined as the percentage of edited nucleotides at all known editing sites. c, Pearson correlations for the editing levels of individual sites between various adult mouse tissues (129S1 strain). d, Numbers of significantly differentially edited sites between various brain parts from 129S1 adult mice (n =2 biological replicates). e, Editing levels of two exemplary sites that are differentially edited between various brain parts from 129S1 adult mice (n =2 biological replicates). f, Pearson correlations for the editing levels of individual sites between various adult mouse tissues (FVB strain). g, Editing levels of two exemplary sites that are differentially edited between various brain parts from FVB adult mice (n =4 biological replicates). h, Comparison of editing levels in the cerebellum and frontal lobe between mice of two different genetic backgrounds (129S1 and FVB). The editing levels of sites that are marked in red differ by more than 10% between the two mouse strains in both cerebellum and frontal lobe. Editing levels were calculated as the average between technical replicates at reproducible sites (P >0.05, Fisher’s exact test, for the comparison of edited and unedited nucleotide counts between technical replicates). i, Predicted RNA secondary structure for part of the NT5DC3 3′UTR that contains an SNP (blue) and an editing site (orange). The editing site in the FVB strain (edited at 63%) is located in a more stable dsRNA stem than the same site in the 129S1 strain (edited at 15%). j, Changes in RNA editing levels during a four-day period of liver regeneration after carbon tetrachloride (CCl₄)-induced injury in the mouse. A total of 262 editing sites were significantly variable from day 0 to day 4 after injury (P <0.2, ANOVA). k-means clustering revealed that the 262 sites can be divided into five distinct groups with different patterns of editing level changes. For each cluster, an exemplary editing site was shown on the right. k, GO analysis of the genes in which editing was dynamically regulated during liver regeneration. During liver injury, hepatocytes undergo necrosis and the surviving hepatocytes proliferate. The enriched GO terms suggest that RNA editing may have an important role during the reparative process of the liver.