Fig. 4
From: NanoMod: a computational tool to detect DNA modifications using Nanopore long-read sequencing data
The average precision and recall rates of NanoMod on 100 simulation data sets. “U test”: Mann–Whitney U test, “T test”: Student’s T test, “KS test”: Kolmogorov-Smirnov test, “Stouffer’s M”: Stouffer’s method and “Fisher’s M”: Fisher’s method. “MeanDif”: modified bases only have the mean difference for signals from non-modified bases, “STDDif”: modified bases only have the difference of standard deviation for signals from non-modified bases, “Mean_STDDif”: modified bases have the difference of the mean and standard deviation for signals from non-modified bases, and “Mean_STDDif_NE”: the simulation of “Mean_STDDif” plus neighborhood effect. Precision was calculated using the number of correctly identified modifications divided by the number of modification predictions at a percentile (i.e., 0.1, 0.25, 0.5, 1, 2, 3, 4 and 5%), and recall was calculated using the number of correctly identified modifications divided by the number of modifications