Supplementary Figure 5: Comparison of DART-Seq to miCLIP.

(a) DART-Seq sites were compared to m⁶A peak regions identified by immunoprecipitation-based m⁶A mapping from three different datasets (PA-m⁶A-Seq¹⁵, m⁶A-Seq⁴, and MeRIP-Seq³). In addition, CIMS/CITS sites from two single-nucleotide resolution miCLIP datasets^5,6 were also compared to m⁶A peak regions from the three datasets. Shown is the overlap between DART-Seq and the two miCLIP datasets reported as the percentage of m⁶A peak sites that overlap. Across all three datasets (PA-m⁶A-Seq, m⁶A-Seq, and MeRIP-Seq), DART-Seq performs similarly to miCLIP in identifying m⁶A sites. (b) m⁶A sites not located within a DRACH consensus motif were taken from the Ke et al. 2015⁶ dataset. DART-Seq C to U editing events which overlapped with CIMS/CITS sites at these non-DRACH m⁶A residues are shown as a percentage of the total number of non-DRACH m⁶A sites. CIMS/CITS sites from Linder et al. 2015⁵ were also compared to these non-DRACH sites, showing a similar degree of overlap as is observed with DART-Seq. (c) Analysis of DART-Seq site clustering suggests minimal promiscuous C to U editing at non-m⁶A-adjacent cytidines. Individual DART-Seq or CIMS/CITS sites were analyzed from DART-Seq data or two miCLIP datasets (Ke et al. 2015 and Linder et al. 2015). For each site within the dataset, the closest C to U editing site (DART-Seq) or CIMS/CITS site (miCLIP) was identified. Shown is the proportion of total sites that are at least 10 nt or 20 nt away from the closest site. All three datasets show a similar distribution of sites, with the majority of sites being at least 10 nt apart.