Quantification of biological network perturbations for mechanistic insight and diagnostics using two-layer causal models

Table 1 Classification of the different methodologies and some of their representatives (see [1], [14] for more exhaustive lists)

Reasoning	Category	Method	Qualitative	Quantitative	Use topology	Interpretability	Diagnostic Sigs	Threshold-free
	ORA	Hypergometric test [15]	√
		GSEA [2]	√	(√)		√		√
	FCS	GSA [52]	√	(√)				√
		PLAGE [16]	√	(√)				√
Forward		CORG [33]				√	√	√
		SPIA [46]	√		√
		NetGen [17]	√		√			√
	PT	NetWalker [18]	√		√	√		√
		SVM-based [35]			√		√	√
	ORA	RCR [21]	√
		Modified GSA [22]	√	√				√
Backward	FCS	MARINa [20]	√					√
		NPA [23]	√	√				√
	PT	TopoNPA	√	√	√	√	√	√

For each represented method, their properties are described. The method described in this study, TopoNPA, is to our knowledge the first backward PT methodology. ORA: Over-Representation Analysis, FCS: Functional Class Scoring, PT: Pathway Topology. (√): with small modification.

ISSN: 1471-2105