Gene regulatory network inference: evaluation and application to ovarian cancer allows the prioritization of drug targets

Table 1 Accuracies of unsupervised and supervised GRNI methods on different datasets

	Unsupervised method		SIRENE
Datasets	Method	AUC	AUC
DREAM3 (knockdown): genes 100, samples 100	MRNET	0.59	0.71
DREAM4 (multifactorial): genes 100, samples 100	GENIE	0.79	0.69
Ovary-normal: genes 2,450, samples 12	RN	0.55	0.62
Ovary-normal: genes 282, samples 12	RN	0.70	0.86

Comparison of accuracies (AUC) of unsupervised and supervised Gene Regulatory Network Inference (GRNI) methods on different datasets. For each dataset, the best-performing unsupervised method was selected for comparison with SIRENE.
AUC, area under the receiver-operating characteristic curve; DREAM, Dialogue for Reverse Engineering Assessments and Methods; GENIE, Gene Network Inference with Ensemble of Trees; MRNET, Minimum Redundancy/Maximum Relevance Networks; RN, Relevance Networks; SIRENE, Supervised Inference of Regulatory Networks.

ISSN: 1756-994X