Exploring glioma heterogeneity through omics networks: from gene network discovery to causal insights and patient stratification

Table 3 Results from spectral clustering based on the variables selected from undirected vs. directed networks; genes covering the top 10% of importance vs. most important gene per module; strength centrality (C_S) vs. eigenvector centrality (C_E) vs. closeness centrality (C_C))

	Undirected networks - Top 10%									Undirected networks - Best per module									Directed networks - Top 10%									Directed networks - Best per module
	C_S			C_E			C_C			C_S			C_E			C_C			C_S			C_E			C_C			C_S			C_E			C_C
	15 features			10 features			29 features			23 features			23 features			26 features			16 features			13 features			30 features			32 features			32 features			29 features
	(set 1)			(set 2)			(set 3)			(set 4)			(set 5)			(set 6)			(set 7)			(set 8)			(set 9)			(set 10)			(set 11)			(set 12)
Astrocytoma	181	24	50	171	21	63	232	21	2	204	26	25	207	31	17	225	26	4	235	18	2	85	46	124	186	41	28	224	12	19	232	21	2	219	8	28
Glioblastoma	32	170	4	45	140	21	14	192	0	16	186	4	12	194	0	10	196	0	66	139	1	15	186	5	13	192	1	27	178	1	28	178	0	32	173	1
Oligodendroglioma	17	4	145	135	4	27	68	1	97	46	4	116	12	9	145	19	3	144	70	2	94	36	11	119	120	10	36	155	2	9	40	2	124	29	0	137
ARI	0.47			0.20			0.57			0.51			0.64			0.72			0.36			0.30			0.35			0.38			0.59			0.58
ASW	0.55			0.42			0.56			0.51			0.63			0.68			0.59			0.43			0.60			0.65			0.49			0.54

ISSN: 1756-0381