A statistical method to incorporate biological knowledge for generating testable novel gene regulatory interactions from microarray experiments

Table 1 Results for the 102-gene set

	LOI.fdr(q)	PCC.fdr(q)	Number of Edges Identified	Published Edges	Same GO BP Annotation	% Published	% Same GO-BP
A	GO-simple	NoRestriction	9147	171	2418	1.9	26.6
B	NoRestriction	PCC(0.1)	1701	37	633	2.2	37.2
	NoRestriction	PCC(0.05)	1377	30	526	2.2	38.2
	NoRestriction	PCC(0.005)	346	12	156	3.5	45.1
C	LOI(.1)	NoRestriction	2312	142	840	6.1	36.3
	LOI(0.05)	NoRestriction	2237	142	809	6.3	36.2
	LOI(0.005)	NoRestriction	1976	136	577	6.9	29.2
D	LOI(.1)	PCC(0.1)	354	26	138	7.3	39.0
	LOI(0.05)	PCC(0.1)	345	26	133	7.5	38.6
	LOI(0.005)	PCC(0.1)	308	26	101	8.4	32.8
E	LOI(.1)	PCC(0.05)	298	22	119	7.4	39.9
	LOI(0.05)	PCC(0.05)	289	22	114	7.6	39.4
	LOI(0.005)	PCC(0.05)	257	22	87	8.6	33.9
F	LOI(.1)	PCC(0.005)	69	10	32	14.5	46.4
	LOI(0.05)	PCC(0.005)	68	10	31	14.7	45.6
	LOI(0.005)	PCC(0.005)	64	10	29	15.6	45.3

A: GO-simple: a pair of genes is considered potentially interacting if one of their GO annotation pairs has been reported.
B: set of interactions obtained by applying only PCC thresholds;
C: set of interactions obtained by applying only LOI thresholds;
D: common set of interactions obtained by applying different LOI thresholds and a fixed PCC.fdr(0.1);
E: common set of interactions obtained by applying different LOI thresholds and a fixed PCC.fdr(0.05);
F: common set of interactions obtained by applying different LOI thresholds and a fixed PCC.fdr(0.005).

ISSN: 1471-2105