Bow-tie networks generally have diverse inputs and outputs with conserved core nodes. It is called “bow-tie” as its pictorial representation resembles bowtie. Fig 1(a): In a three layer feed forward neural networks, a hidden layer (a middle layer) provides generalization capability and the neural network can function as classifier of diverse inputs mapping into diverse output patterns. It is essential that numbers of nodes in the hidden layer is smaller than numbers of nodes in inputs and output layers in order to achieve high level of generalization. This is an example of 6∶3∶6 feed forward network with 6 input nodes, 3 hidden nodes, and 6 output layer nodes. Fig 1(b): In signaling networks, receptors corresponds nodes in an input layer, molecules such as cAMP and calcium corresponds nodes in a hidden layer, and transcription factors corresponds to nodes in an output layer (Left). Looking at this from a stimuli-response viewpoint, it shall be viewed as a process that diverse ligands activate variety of receptors forming distinct activation patterns and results in patterns of activations at transcriptional level (Right). An intermediate layer (core nodes) that corresponds to the hidden layer in the feed forward neural network shall provide generalization capability to the signaling network. In GPCR pathway, cAMP and calcium are key molecules constituting this layer. Thus, diverse stimuli are classified into several groups that have similar calcium and cAMP elevation. Ligands that are classified into the same group shall activate similar subset of genes, hence invoking similar physiological responses, if generalization is actually taking place.