Skip to main content
. 2019 Mar 26;17:454–462. doi: 10.1016/j.csbj.2019.03.013

Fig. 3.

Fig. 3

Predictive quality for the predictors of MoRF regions measured on the TEST419 dataset (for which MoRF annotations are based on putative disorder) and TEST53/TEST45 (for which MoRF annotations are based on experimentally verified disorder). Panel A shows the AUC values. Panel B gives the values of sensitivity (measured for FRP = 0.1 and shown using circles) and specificity (measured for TRP = 0.5 and shown with crosses). The results were taken from the original publications. All predictors were developed using the same training dataset (TRAIN419) that shares low sequence similarity (<30%) with these test datasets.