Fig. 4. An example of a Bayesian Network model for four genes. If we assume that Gene 1 activates Gene 2, then we can construct a conditional probability table (shown at the right) that captures our observations of the state of Gene 1 when we observe Gene 2 to be upregulated. Here the values for Gene 1 of –1, 0, and +1 represent states where Gene 1 is downregulated, unchanging or upregulated, respectively.