Table 1.

The virus-human PPI realistic benchmark datasets’ statistics

	Training data				Testing data
	$|{\mathbf{E}}^{+}|$	$|{\mathbf{E}}^{-}|$	$|{\mathbf{V}}^{\mathbf{h}}|$	$|{\mathbf{V}}^{\mathbf{v}}|$	$|{\mathbf{E}}^{+}|$	$|{\mathbf{E}}^{-}|$	$|{\mathbf{V}}^{\mathbf{h}}|$	$|{\mathbf{V}}^{\mathbf{v}}|$
Novel H1N1	10,858	Varies	7636	641	381	Varies	622	11
Novel Ebola	11,341	Varies	7816	659	150	Varies	290	3
Zhou’s H1N1	10,858	10,858	7636	641	381	381	622	11
Zhou’s Ebola	11,341	11,341	7816	659	150	150	290	3
2697049	24,698	246,980	16,638	1066	278	448,651	16,627	27
333761	23,892	238,920	16,638	1070	534	132,482	16,627	8
2043570	24,372	243,720	16,638	1085	309	66,199	16,627	4
644788	24,825	248,250	16,638	1090	54	33,200	16,627	2

$|E^{+}|$ and $|E^{-}|$ refer to the number of positive and negative interactions, respectively. $|V^h|$ and $|V^v|$ are the number of human proteins and virus proteins