Table 1. Open in new tab Performance...

Table 1.

Performance comparison of SVM, RF, D-SCRIPT, Topsy-Turvy, Alphafold3, PIPR, and ProNEP on the independent test dataset.^a

Model	Accuracy	Precision	Sensitivity	F1-score	AUROC	AUPRC
SVM	0.781 ± 0.024	0.157 ± 0.028	0.323 ± 0.037	0.277 ± 0.168	0.617 ± 0.065	0.176 ± 0.057
RF	0.806 ± 0.075	0.749 ± 0.086	0.610 ± 0.122	0.716 ± 0.121	0.783 ± 0.127	0.483 ± 0.117
D-SCRIPT	0.609 ± 0.073	0.324 ± 0.233	0.719 ± 0.252	0.335 ± 0.056	0.686 ± 0.065	0.306 ± 0.093
Topsy-Turvy	0.672 ± 0.117	0.225 ± 0.128	0.546 ± 0.329	0.221 ± 0.046	0.64 ± 0.046	0.359 ± 0.117
Alphafold3	0.185	0.0952	0.9189	0.1726	0.5688	0.1185
PIPR	0.912 ± 0.008	0.569 ± 0.032	0.594 ± 0.026	0.581 ± 0.026	0.879 ± 0.016	0.678 ± 0.004
ProNEP	0.914 ± 0.028	0.945 ± 0.035	0.915 ± 0.036	0.928 ± 0.022	0.966 ± 0.012	0.747 ± 0.054

Model	Accuracy	Precision	Sensitivity	F1-score	AUROC	AUPRC
SVM	0.781 ± 0.024	0.157 ± 0.028	0.323 ± 0.037	0.277 ± 0.168	0.617 ± 0.065	0.176 ± 0.057
RF	0.806 ± 0.075	0.749 ± 0.086	0.610 ± 0.122	0.716 ± 0.121	0.783 ± 0.127	0.483 ± 0.117
D-SCRIPT	0.609 ± 0.073	0.324 ± 0.233	0.719 ± 0.252	0.335 ± 0.056	0.686 ± 0.065	0.306 ± 0.093
Topsy-Turvy	0.672 ± 0.117	0.225 ± 0.128	0.546 ± 0.329	0.221 ± 0.046	0.64 ± 0.046	0.359 ± 0.117
Alphafold3	0.185	0.0952	0.9189	0.1726	0.5688	0.1185
PIPR	0.912 ± 0.008	0.569 ± 0.032	0.594 ± 0.026	0.581 ± 0.026	0.879 ± 0.016	0.678 ± 0.004
ProNEP	0.914 ± 0.028	0.945 ± 0.035	0.915 ± 0.036	0.928 ± 0.022	0.966 ± 0.012	0.747 ± 0.054

ProNEP was trained 10 times with 10 different negative samples, while the other models, excluding Alphafold3, were trained 6 times. All models, except Alphafold3, used a threshold that maximized the F1 score.

This Feature Is Available To Subscribers Only