Transcriptome RNA hyperediting (HE) clusters in zebrafish.

A. Out of the overall 12 possible mismatches between RNA and DNA, there was a cluster enrichment of A-to-G transitions (93%) compared with the other 11 mismatch possibilities (7%). Web-logo diagrams show the abundance of each nucleotide located 1 bp upstream (position -1) and 1 bp downstream (position +1) of every A/G mismatch found in zebrafish. Top panels show the results for both -1 and +1 positions in the zebrafish exons (top right) and whole transcriptome (top left). Consistent with the established motif, ‘G’ is the least represented nucleotide, with 8.4% in whole transcriptome and 11.4% in exons. Bottom panels show the results for both -1 and +1 positions surrounding the 42,500 editing sites comprising the RADAR dataset in humans (bottom left) and mice (bottom right). In position -1, similar to the case of zebrafish, ‘G’ is considerably under-representation with 8.4% and 4.4% in humans and mice, respectively. B. The distribution and number of A-to-I RNA hyperediting (HE) sites. The top chart represents all detected DNA-RNA mismatches. A-to-G mismatches are the majority (93%) of all mismatches. Middle and bottom charts show the genomic location of the detected HE clusters. C. Most of the detected RNA editing sites were found in repeats. Comparison between the distribution of the total and edited repeat families in the zebrafish genome showed an enrichment of the hAT family DNA repeats. While the hAT family occupies only 8% of total repeats in the entire zebrafish genome, it holds 26% of the total cluster containing sequences. D. mfold analysis of RNA secondary structure performed on the two most prominent DNA repeats (ANGEL and TDR19), which are members of the hAT family and account for over 11% of all sites detected. Structure analysis shows a long-stemmed dsRNA structure with palindrome traits that enable Adar binding and, consequently, RNA editing. Color code represents the strength of the nucleotide connection.A. 在RNA与DNA间共计12种可能的错配类型中，A-to-G转换的簇状富集占比达93%，其余11种错配类型合计仅占7%。WebLogo序列标识图展示了斑马鱼所有A/G错配位点上下游各1个碱基对（bp）位置的核苷酸丰度分布，其中上游为-1位、下游为+1位。上方面板展示了斑马鱼外显子（右上）及全转录组（左上）中-1与+1位的分析结果。与已报道的基序（motif）一致，鸟嘌呤（G）是占比最低的核苷酸：全转录组中占比为8.4%，外显子中占比为11.4%。下方面板则展示了人类（左下）与小鼠RADAR数据集（RADAR dataset）中42500个编辑位点上下游-1及+1位的分析结果。在-1位，与斑马鱼的情况一致，鸟嘌呤（G）的占比显著偏低：人类与小鼠中分别为8.4%与4.4%。 B. 本部分展示A-to-I RNA超编辑（A-to-I RNA hyperediting, HE）位点的分布与数量。顶部图表涵盖所有检测到的DNA-RNA错配，其中A-to-G错配占绝大多数（93%）；中部与底部图表则展示了检测到的超编辑簇的基因组定位。 C. 绝大多数检测到的RNA编辑位点位于重复序列中。对斑马鱼基因组中总重复序列家族与已编辑重复序列家族的分布进行比较后发现，hAT家族DNA重复序列存在富集现象：hAT家族仅占斑马鱼全基因组总重复序列的8%，却承载了26%的含编辑簇序列。 D. 针对属于hAT家族且占所有已检测编辑位点逾11%的两个典型DNA重复序列（ANGEL与TDR19），开展了mfold RNA二级结构分析。结构分析结果显示，其存在具备回文特征的长茎双链RNA（double-stranded RNA, dsRNA）结构，可结合ADAR腺苷脱氨酶（adenosine deaminase acting on RNA）并介导RNA编辑。色标代表核苷酸间碱基连接的强度。