Construction of gene regulatory networks using biclustering and bayesian networks

Table 7 Performance comparison of the biclustering networks.

Methods	EdgeCount	TP	FP	TN	FN	AUROC	AUPR
Gold	2194	2194	0	400396	0	1	1
ALL	5440	94	5346	395050	2100	0.5148	0.0073
SAMBA	1611	46	1565	398831	2148	0.5085	0.0072
ISA	2558	56	2502	397894	2138	0.5097	0.0067
OPSM	220	12	208	400188	2182	0.5025	0.0067
Friedman	947	22	925	399471	2172	0.5039	0.0065
CMSBE	735	20	715	399681	2174	0.5037	0.0063
K-means	380	13	367	400029	2181	0.5025	0.0061
Bivisu	1515	13	1502	398894	2181	0.5011	0.0055
CC	590	3	587	399809	2191	0.5000	0.0054

Performances of biclustering networks are compared with the Friedman network and gold network. EdgeCount: the number of network edges; TP: number of true positive edges; TN: number of true negative edges; FP: number of false negative edges; AUROC: area under ROC curve; AUPR: area under precision recall curve.

ISSN: 1742-4682