Novel editing sites have an ADAR sequence motif and are enriched in regions with dsRNA character. (A) Euler diagram depicting the number of significant editing sites detected by the GATK or hyperediting pipeline. Cold-enriched sites were defined as the union of the two sets of significant sites. (B) Euler diagram illustrating the set of constitutively edited sites defined as the nonsignificant editing sites (FDR > 0.5) identified by both the GATK and the hyperediting pipeline. Only sites identified by both methods were selected to reduce the likelihood of misclassifying a SNP as an editing site. (C) Editing index for cold-enriched and constitutively edited sites. (D) Heatmap depicting editing sites conserved in either mouse or human, ordered by hierarchical clustering. Predicted locations of these sites are based on annotations in the RADAR database. (E) Sequence logo of sequences surrounding all cold-enriched or constitutive editing sites (±5 nt) or editing sites identified in mouse in the RADAR database. A set of randomly positioned sites was also generated by randomly selecting an A nucleotide position from transcripts containing editing sites. (F) Quantification of dsRNA character of each editing site via blastn by aligning the reverse complement of ±100 nt regions surrounding editing sites, to a ±2 kb region surrounding the editing site or to the end of the transcript. Sites with an e-value <0.1 with alignment over the editing site are considered paired dsRNA. As a negative control, the nonreverse complement, denoted as (−), of each ±100 nt region was also quantified and the number of second best alignments are shown.
