Figure 3.
Datasets and models. (a) Types of connections in real-world networks: (i) interactions between proteins, (ii) connections between brain regions in the macaque, (iii) connections using chemical synapses, in C. elegans, and (iv) flight connections between airports. Note that connections for (ii) and (iii) may be unidirectional. (b) Growth models leading to highly connected nodes. PA, preferential attachment [6]: new nodes (red) preferentially connect to nodes with higher degrees. DD, duplication–divergence model: at each step, a random node (light red) is duplicated (red) together with its links. NL, nonlinear growth: the number of new nodes that are added at each step increases nonlinearly over time. New nodes project to already present nodes (blue) establishing on average a connections (NLA) or link to each existing node with a probability p (NLP). (c) Exemplary distributions for node degree k (inset: log–log plot) for networks from linear (red) and nonlinear growth (blue). Linear growth corresponds to the scenario where the network size increases linearly, i.e. only one node is added at each time step. Shaded areas show the standard deviation around the mean degree (dashed line). Nonlinear growth yields a wider distribution with more hubs.