A multitask transfer learning framework for the prediction of virus-human protein–protein interactions

BMC Bioinformatics

Table 6 Comparison with methods based on hand-crafted features

Dataset	Model	AUC	AP	Precision	Recall	F1
Zhou’s H1N1	Denovo	0.8656	0.8619	77.75	77.95	77.85
	Generalized	0.8600	0.8606	76.96	77.17	77.06
	Hybrid	0.937	–	–	–	–
	MTT	0.9461	0.9589	86.28	86.51	86.40
Zhou’s Ebola	Denovo	0.8864	0.8366	83.44	84.00	83.72
	Generalized	0.9154	0.9078	84.77	85.33	85.05
	MTT	0.9680	0.9766	90.93	91.53	91.23
Denovo_slim	Denovo	0.8701	0.8631	81.92	82.12	82.02
	Generalized	0.8891	0.8851	84.74	84.94	84.84
	MTT	0.9221	0.9324	83.92	84.12	84.02
Barman	Denovo	0.8217	0.8415	74.60	74.98	74.79
	Generalized	0.8214	0.8458	74.90	75.27	75.08
	MTT	0.9804	0.9802	93.53	94.05	93.79
Bacillus	Denovo	0.9843	0.9650	94.80	94.83	94.83
	Generalized	0.9833	0.9668	95.75	95.78	95.76
	MTT	0.9997	0.9992	98.75	98.78	98.76
Yersina	Denovo	0.9712	0.9302	93.14	93.16	93.15
	Generalized	0.9758	0.9362	94.01	94.03	94.02
	MTT	0.9988	0.9971	97.32	97.34	97.32
Franci	Denovo	0.9782	0.9584	95.55	95.62	95.58
	Generalized	0.9799	0.9565	95.84	95.91	95.88
	MTT	0.9998	0.9996	98.95	99.03	98.99

ISSN: 1471-2105