Figure 2

TopoNPA workflow takes as input gene log ₂ -fold changes and a fixed network structure. a) The measured gene l o g₂-fold changes are used to infer the differential backbone values for which no direct measurements are available. This step relies on the backward assumption. The differential values obtained on the backbone are summarized in a single number quantifying how strong they are and how well those values agree with the signed graph structure of the backbone. b) Once the perturbation is quantified, three statistics are computed to assess the significance of TopoNPA with respect to the experimental error and its specificity to the given two-layer network structure.