Table 1.

Performance of our deep learning architecture (PSSM+CNN+MLP) using 5-fold cross-validation^a

Human–virus PPI dataset	Accuracy (%)	Precision (%)	Sensitivity (%)	Specificity (%)	F1-score (%)	AUPRC
Human–HIV	98.65	95.16	89.72	99.54	92.36	0.974
Human–Herpes	95.26	77.11	68.10	97.98	72.33	0.768
Human–Papilloma	95.98	82.70	70.48	98.53	76.10	0.818
Human–Influenza	96.10	84.22	70.30	98.68	76.63	0.834
Human–Hepatitis	93.43	69.27	49.77	97.79	57.92	0.636
Human–Dengue	93.29	70.02	45.85	98.04	55.41	0.605
Human–Zika	95.41	85.17	59.94	98.96	70.36	0.746
Human–SARS-CoV-2	90.64	45.81	16.37	98.06	24.12	0.329

^aThe definitions of performance assessment metrics are available in Supplementary Materials.