Figure 3

Performance of different methods in the analysis of noise-disturbed data. The performance of methods in the analysis of PFMs with 75% noise added. The ROC curves show the sensitivity (y-axis) and false positive rate, i.e. 1- specificity, (x-axis). Different combinations of the scoring functions of Matlign are shown in grey, with the best one highlighted in black. T-Reg (green) was run using default parameters and with the given orientation of the cis-elements [9]; MatCompare (blue) was run as recommended by the authors [11]. The performance of CompareACE/Pearson (brown) and YSRA (red) was simulated by implementing their corresponding distance functions in Matlign.