Figure 1. Network segregation for two attractor states ().

Every edge that connects a similarity node to a differential node or a differential node to a similarity node transmits no signal. This means that the signaling in the right network shown above is identical to that of the left network. Because the goal is to leave normal cells unaltered while damaging cancer cells as much as possible, all similarity nodes can be safely ignored, and searches and simulations only need to be done on the differential subnetwork.