Skip to main content
. 2021 Jul 21;25:515–523. doi: 10.1016/j.omtn.2021.07.008

Figure 2.

Figure 2

Quantification and detection of endogenous RNA editing by MultiEditR

(A) Workflow for generation of sequencing data for three-way comparison of MultiEditR, RNA-seq, and Amplicon-seq. RNA-seq from WT cell lines was compared to RNA-seq from paired APOBEC1 or ADAR1 KO cell lines to identify putative editing sites, which were PCR amplified from WT cell line cDNA. Amplicons were then subjected to Sanger sequencing and NGS deep sequencing in parallel and analyzed by MultiEditR or REDItools, respectively. (B–E) Comparison of MultiEditR and RNA-seq measurements of editing to high coverage Amplicon-seq measurements. All data are filtered on between 1% and 99% editing as measured from Amplicon-seq data. Coefficients of determination (R2) represent regression to the identity line. Dot size is proportional to read coverage at the base of interest in the RNA-seq data. M = mean, SD = standard deviation. p values (P) were calculated by Student’s one-sample two-tailed t test. Boxplot center lines represent the median, box limits represent the upper and lower quartiles, and whiskers define the 1.5× interquartile range. (F) Kernel plots of the inaccuracy of MultiEditR and RNA-seq relative to Amplicon-seq. Plots are faceted by 1%, 5%, or 10% editing inclusion cutoffs based on the editing measured in Amplicon-seq. (G) Confusion matrix that defines the sensitivity and specificity of the edit detected by MultiEditR or REDItools, against the true state of editing, defined by Amplicon-seq. (H) ROC curves of MultiEditR and REDItools at 1%, 5%, and 10% editing cutoffs, where an Amplicon-seq event is a true edit. MultiEditR is shown as a solid curve, while REDItools is shown as a dotted curve. Crossed points along the curves represent the closest geometric distance of the curve to the upper-left error-free point at 0% and 100%. (I) ROC analysis metrics for MultiEditR and REDItools at editing thresholds from 0.1% to 25% editing inclusion cutoff. Sensitivity and specificity are defined in (G); area under the curve (AUC) represents the integral of the ROC curve from 0% to 100%.