Skip to main content

Table 3 Performance comparison with existing methods on the benchmark test dataset

From: Co-AMPpred for in silico-aided predictions of antimicrobial peptides by integrating composition-based features

Method Acc AUROC AUCPR Kappa Sen Spe MCC References
iAMP-2L 65.4% 0.318 82.9% 47.9% 0.329 Xiao et al. [25]
iAMPpred 70.7% 0.415 80.8% 60.6% 0.424 Meher et al. [26]
AmPEP 68.0% 0.751 0.686 0.362 93.6% 42.5% 0.421 Bhadra et al. [27]
AMP Scanner DNN 73.4% 0.806 0.777 0.468 80.8% 65.9% 0.473 Veltri et al. [28]
RF-AmPEP30 77.1% 0.854 0.868 0.543 77.6% 76.6% 0.542 Yan et al. [32]
Deep-AmPEP30 77.1% 0.853 0.853 0.543 76.6% 77.7% 0.543 Yan et al. [32]
Co-AMPpred 80.8% 0.871 0.890 0.606 79.7% 81.9% 0.606 This study
Co-AMPpred70 78.6% 0.861 0.860 0.553 80.9% 74.5% 0.554 This study
Co-AMPpred80 76.6% 0.851 0.840 0.532 78.7% 74.5% 0.532 This study
Co-AMPpred90 70.2% 0.843 0.860 0.404 89.4% 51.1% 0.438 This study
  1. The given data in bold font indicates the top performance of the model on the test dataset
  2. Acc., accuracy; AUROC, area under the receiver operating characteristics curve; AUCPR, area under the precision-recall curve; Sen., sensitivity; Spe., specificity; MCC, Matthew's correlation coefficient; SD, standard deviation