Fig. 2

Performance on simulated iCLIP-seq data. Precision vs number of true positive crosslink sites (top) and vs number of true positive binding regions, i.e., regions with at least one correctly called crosslink site (bottom) for three different simulation settings. The characteristics of each simulation are reported below the plots. All strategies were applied using different sensitivity settings, i.e., different p value or score thresholds. The leftmost point of each curve corresponds to the number of true positives associated with the lowest p value or the highest score the strategy can report. The curves of Piranha + CITS ^sc overlap with those of CITS