Performance comparison of PROSPER with PoPS and SitePrediction for predicting cleavage sites for selected enzymes.

a“-” denotes that the prediction result at this specificity level is not available by this tool.Specificity (Sp) levels were set as close as possible to one another among the different tools. For PROSPER, 5-fold cross-validation ± feature selection and self-consistency tests were performed to compare to PoPS and SitePrediction. The 3 different types of PROSPER models built are named PROSPER5CV, PROSPERselect and PROSPERSelf, respectively. Sensitivity (Sn) values at different Sp levels were compared. The best prediction performance at each Specificity level is highlighted in bold.