Fig. 3

Distribution of raw read counts and structure scores obtained using reactIDR at stem and loop loci. a Violin plots, showing the ascending rank orders of the four indices of chemical footprint enrichment that can be computed for four rRNAs from case and control samples separately: Count, raw read count; Ratio, read count normalized by the reads passing through the position; irreproducible discovery rate (IDR); and reactIDR. With the increase in the rank of the structure score obtained from the case samples, the likelihood of the nucleotide being in the loop structure increases. b Read count distribution of reproducible and irreproducible groups, classified into top 25% and bottom 25% rankings, based on the IDR (left) and reactIDR reactivity indicator(right) criteria. c Receiver operating characteristic (ROC) curve, showing stem/loop structure prediction of the test set which is composed of 5S, 5.8S, and 28S rRNAs based on the structure scores obtained using six scoring methods, for the in vivo icSHAPE dataset