Table 1. Comparison of the global properties of the four predicted networks.

Network type	Nodes	Edges	VV seeds leading to predictions	Connected components (CC)	Size of largest CC	Transitivity	Predictions
Sequence network	6349	27807	19 (12)	1155	625	0.69	100 (61)
Core PPI [14]	3459	24348	30 (21)	111	2995	0.85	79 (54)
Co-expression	3654	33272	18 (13)	159	3239	0.42	47 (14)
Integrated	9521	80997	60 (50)	439	8364	0.52	215 (120)

Global properties of the three constituent networks and the integrated network. The sequence similarity network excludes nodes with no edges (‘orphan’ proteins with no sequence similarity matches). Column 4 is the number of verified virulence (VV) seeds involved in the predictions, using the rule that a node must be connected to at least 2 seeds to be a prediction (in brackets are the corresponding numbers if we require connection to at least 3 seeds); Column 8 gives the number of predictions (in brackets are the corresponding prediction counts if we require a more stringent rule i. e. a node must be connected to at least 3 seeds to be a prediction).