Fig. 6

Binary classification of missense SCN9A-gene mutation sites based on their distance from SF’s critical point. a ROC curve constructed from data of distances between mutation sites and SF’s critical-point (for construction of data set see Additional file 1: S9b). Optimal threshold value \(\sim\)5.8 Å corresponds to specificity and sensitivity values of 0.812 and 0.777, respectively. Area under ROC curve is 0.824. b Visualization of ROC curve data. Optimal threshold value \(\sim\)5.8 Å is marked with black dashed line. ROC curve is constructed in R [73] by using the pROC package [93]. Two sets of missense SCN9A-gene mutation sites are employed; a pain-related set containing IEM, PPD and SFN mutation sites, and a neutral set containing mutation sites which are not expected to associate with pain disease phenotypes (Additional file 1: S8)