Skip to main content
. 2012 Nov 29;7(11):e50300. doi: 10.1371/journal.pone.0050300

Table 3. Performance comparison of PROSPER with PoPS and SitePrediction for predicting cleavage sites for selected enzymes.

Caspase-1 MMP-2
PoPS PoPS
Sp (%) 99.9 99.8 99.5 99.0 -a Sp (%) 99.9 99.8 99.5 99.0 -
Sn (%) 14.0 34.0 48.0 72.0 - Sn (%) 4.2 5.8 15.4 27.2 -
SitePrediction SitePrediction
Sp (%) 99.9 99.8 99.5 99.0 98.0 Sp (%) 99.9 99.8 99.5 99.0 98.0
Sn (%) 60.0 68.0 72.0 86.0 92.0 Sn (%) 6.7 12.4 22.2 36.4 50.8
PROSPER5CV PROSPER5CV
Sp (%) 99.9 99.8 99.5 98.9 98.0 Sp (%) 99.8 99.7 99.4 99.0 98.1
Sn (%) 12.0 14.0 36.0 56.0 80.0 Sn (%) 8.8 15.3 23.9 31.0 43.8
PROSPERSelect PROSPERSelect
Sp (%) 99.9 99.8 99.4 99.0 98.0 Sp (%) 99.9 99.7 99.5 99.0 98.1
Sn (%) 22.0 40.0 66.0 80.0 88.0 Sn (%) 10.3 17.9 23.0 32.1 39.7
PROSPERSelf PROSPERSelf
Sp (%) 99.9 99.8 99.5 99.0 98.1 Sp (%) 99.8 99.7 99.4 99.1 98.0
Sn (%) 10.0 36.0 88.0 94.0 98.0 Sn (%) 8.1 18.0 36.0 48.2 67.5
Caspase-3 MMP-9
PoPS PoPS
Sp (%) 99.9 99.8 99.5 99.0 - Sp (%) 99.9 99.8 99.5 99.2 -
Sn (%) 42.6 50.4 72.3 86.6 - Sn (%) 2.8 9.5 12.8 22.3 -
SitePrediction SitePrediction
Sp (%) 99.9 99.8 99.5 99.0 98.0 Sp (%) 99.9 99.8 99.5 99.0 98.0
Sn (%) 52.5 62.7 81.6 91.2 97.4 Sn (%) 2.4 7.6 14.7 21.3 41.7
PROSPER5CV PROSPER5CV
Sp (%) 99.9 99.7 99.5 99.0 98.0 Sp (%) 99.9 99.8 99.5 99.0 98.0
Sn (%) 28.6 56.3 68.2 81.9 93.6 Sn (%) 2.4 6.2 12.3 20.4 29.9
PROSPERSelect PROSPERSelect
Sp (%) 99.8 99.7 99.5 99.0 98.0 Sp (%) 99.9 99.8 99.5 99.0 98.0
Sn (%) 51.3 66.2 76.7 86.0 93.6 Sn (%) 3.3 6.2 10.4 17.1 28.4
PROSPERSelf PROSPERSelf
Sp (%) 99.9 99.8 99.5 99.0 98.0 Sp (%) 99.9 99.8 99.5 98.9 98.0
Sn (%) 30.0 76.1 92.4 97.4 99.4 Sn (%) 3.8 9.5 28.4 48.8 61.6
Caspase-7 GrB (human)
PoPS PoPS
Sp (%) 99.9 99.8 99.5 99.0 - Sp (%) 99.9 99.8 99.5 99.0 -
Sn (%) 56.2 65.2 79.8 93.3 - Sn (%) 17.4 23.6 34.1 43.8 -
SitePrediction SitePrediction
Sp (%) 99.9 99.8 99.5 99.0 98.0 Sp (%) - - - - -
Sn (%) 57.3 66.3 76.4 89.9 96.6 Sn (%) - - - - -
PROSPER5CV PROSPER5CV
Sp (%) 99.9 99.8 99.5 99.0 98.0 Sp (%) 99.9 99.8 99.6 99.0 98.0
Sn (%) 33.7 49.4 75.3 85.4 87.6 Sn (%) 19.9 37.0 53.3 67.8 83.0
PROSPERSelect PROSPERSelect
Sp (%) 99.9 99.8 99.5 99.0 98.0 Sp (%) 99.9 99.8 99.5 99.0 98.0
Sn (%) 40.4 58.4 84.3 87.6 89.9 Sn (%) 12.7 24.3 35.9 55.1 71.0
PROSPERSelf PROSPERSelf
Sp (%) 100 99.8 99.5 99.1 98.1 Sp (%) 99.9 99.8 99.5 99.0 98.0
Sn (%) 61.8 97.8 100 100 100 Sn (%) 23.9 47.5 83.7 93.8 98.6
Caspase-6 GrB (mouse)
PoPS PoPS
Sp (%) 99.9 99.8 99.5 99.0 - Sp (%) 99.9 99.8 99.5 99.0 -
Sn (%) 20.4 23.4 44.9 68.3 - Sn (%) 14.9 20.5 33.6 50.8 -
SitePrediction SitePrediction
Sp (%) 99.9 99.8 99.5 99.0 98.0 Sp (%) - - - - -
Sn (%) 20.4 37.1 80.2 92.8 95.2 Sn (%) - - - - -
PROSPER5CV PROSPER5CV
Sp (%) 99.9 99.8 99.4 99.0 98.0 Sp (%) 99.9 99.8 99.5 98.9 97.9
Sn (%) 15.0 25.8 53.3 66.5 83.8 Sn (%) 21.4 34.4 45.5 55.8 73.4
PROSPERSelect PROSPERSelect
Sp (%) 99.9 99.8 99.5 98.9 98.1 Sp (%) 99.9 99.8 99.5 99.1 98.1
Sn (%) 16.8 29.9 55.7 77.8 92.8 Sn (%) 18.2 29.2 46.8 63.6 79.2
PROSPERSelf PROSPERSelf
Sp (%) 99.9 99.8 99.5 99.0 98.1 Sp (%) 99.9 99.8 99.5 99.0 98.0
Sn (%) 15.6 50.9 68.3 84.4 98.8 Sn (%) 27.9 54.6 77.9 92.9 98.0
Caspase-8
PoPS
Sp (%) 99.9 99.8 99.5 99.0 -
Sn (%) 48.3 60.4 77.6 91.4 -
SitePrediction
Sp (%) 99.9 99.8 99.5 99.0 -
Sn (%) 48.3 69.0 89.7 96.6 -
PROSPER5CV
Sp (%) 99.9 99.8 99.5 99.0 98.0
Sn (%) 22.4 36.2 46.6 69.0 77.6
PROSPERSelect
Sp (%) 99.9 99.8 99.6 99.0 98.0
Sn (%) 25.9 51.2 62.1 84.5 91.4
PROSPERSelf
Sp (%) 99.9 99.8 99.5 99.0 98.0
Sn (%) 13.8 41.4 91.4 98.3 98.3
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.