Figure 5.

A hypothetical gene network and corresponding Markov random graph. (a) A small hypothetical gene network with perturbation in two datasets D_A and D_B. The genes in the two datasets interact through identical network, although they assume different states. The circle g₀ represents the abstraction of the external perturbation. Rectangles denote genes. → implies activation and ⊣ implies inhibition. The potential effect of metagene to all other genes is indicated by dotted arrows from the metagene to all the other genes. For example, g₁ is primarily affected in D_A, but not affected in D_B. g₂ is primarily affected in both the datasets. g₃ is secondarily affected in both D_A and D_B. (b) The Markov Random Field graph constructed based on the small hypothetical gene network in (a). The numbers in the parenthesis are the expected assignments to the variables based on the states of the genes in (a). Nodes with dotted boundaries indicate that those nodes are required for completeness of the model, however the corresponding interactions do not exist.