BMC Bioinformatics

Table 1 Netpath consistency scores in breast cancer

From: DART: Denoising Algorithm based on Relevance network Topology improves molecular pathway activity inference

		ER-				ER+
Pathway	nG	nE	fE	fconsE	Pval	nE	fE	fconsE	Pval
a6b4	27	33	0.09	0.64	0.05	94	0.27	0.53	0.22
AR	511	6164	0.05	0.55	< 0.001	22486	0.17	0.54	< 0.001
BCellReceptor	396	5324	0.07	0.51	0.10	16503	0.21	0.50	0.17
EGFR1	236	2256	0.08	0.56	< 0.001	5896	0.21	0.54	< 0.001
IL1	231	1926	0.07	0.60	< 0.001	5458	0.21	0.56	< 0.001
IL2	722	18836	0.07	0.54	< 0.001	52916	0.20	0.52	< 0.001
IL3	49	99	0.08	0.64	< 0.001	257	0.22	0.57	0.01
IL4	292	3463	0.08	0.54	< 0.001	9531	0.22	0.53	< 0.001
IL5	167	1109	0.08	0.75	< 0.001	3330	0.24	0.64	< 0.001
IL6	104	250	0.05	0.62	< 0.001	1037	0.19	0.62	< 0.001
IL7	62	189	0.10	0.63	< 0.001	353	0.19	0.59	< 0.001
IL9	24	12	0.04	0.92	< 0.001	47	0.17	0.83	< 0.001
KitReceptor	70	115	0.05	0.79	< 0.001	477	0.20	0.60	< 0.001
Notch	92	313	0.07	0.53	0.13	876	0.21	0.51	0.24
RANKL	69	147	0.06	0.62	< 0.001	394	0.17	0.53	0.14
TCellReceptor	561	11587	0.07	0.59	< 0.001	31820	0.20	0.55	< 0.001
TGFBReceptor	993	21396	0.04	0.53	< 0.001	92352	0.19	0.51	< 0.001
TNFA	801	11226	0.04	0.60	< 0.001	54534	0.17	0.53	< 0.001

For both the ER- and ER+ breast cancer data set [24] and for a number of important Netpath cancer signalling and immune signalling pathways (see Methods), we list some of the network properties of the inferred relevance expression correlation networks. For each molecular pathway we give the number of genes of the pathway present in the expression matrix (nG), the number and fraction of edges (i.e significant pairwise correlations between genes) (nE & fE), the fraction of edges that are consistent with the prior information (fconsE) and the corresponding p-value of significance (Pval). P-values were estimated using 1000 permutations.

Back to article page

ISSN: 1471-2105

Contact us

General enquiries: journalsubmissions@springernature.com