A multilevel layout algorithm for visualizing physical and genetic interaction networks, with emphasis on their modular organization

Table 1 Interaction networks used in the study

Network name	Ref	Type	Screen	Nodes	Edges	D	MND	MND_SD	MCC
Ito-Core	[42]	PI	PPI	426	568	0.006	2.667	3.919	0.093
VonMering	[49]	PI	PPI	573	2097	0.013	7.319	9.017	0.450
Schwikowski	[50]	PI	PPI	1297	1862	0.002	2.871	3.109	0.125
Y2H-CCSB	[43]	PI	PPI	964	1598	0.003	3.315	5.456	0.095
Y2H-Union	[43]	PI	PPI	1647	2682	0.002	3.257	5.334	0.086
AP/MS-Combined	[45]	PI	CCA	1004	8319	0.017	16.57	18.63	0.648
LC-Multiple	[48]	MT	LCI	1213	2621	0.004	4.322	4.533	0.337
Secretory-Map	[53]	GI	E-MAP	409	4175	0.050	20.42	23.82	0.251
Chromosome-Map	[54]	GI	E-MAP	735	17,185	0.064	46.76	43.61	0.233
Costanzo	[55]	GI	SGA	4319	74,984	0.007	29.96	41.86	0.062
Costanzo-Stringent	[55]	GI	SGA	3811	35,924	0.004	16.07	22.92	0.046

Network types: PI, physical interactions; GI, genetic interactions; MT, mixed type. Screening methods: PPI, protein-protein interaction screen; CCA, protein co-complex association mapping; LCI, literate-curated interactions; E-MAP, epistatic miniarray profiling; SGA, synthetic genetic array mapping. Topological parameters: D, density; MND, mean node degree; MND_SD, standard deviation of MND; MCC, mean clustering coefficient of the network. Costanzo-Stringent sub-network was constructed using the interaction score cut-offs ε < -0.17 or ε > 0.21. In each network, we extracted the largest connected component to be used in the evaluations.