Each of the 10 randomized networks that we built had the same number of nodes, connections and connectivity distribution as the MADS network (see Methods).
We compared subgraph counts of this network to that of 10 of randomized networks that preserved the number of nodes, connections, and connectivity distribution.
This is unlikely, since we compare each network with randomized versions with the same number of nodes, connections, and connectivity distribution.
To characterize the overall network topology of the 5' and 3' halves, we compared the 2 networks by measuring the distribution of the number of connections per node (connectivity distribution).
Firstly, it has a scale-free Power Law connectivity distribution; a few highly connected nodes or 'hubs', with many poorly connected nodes.
Furthermore, the connectivity distribution of the number of inter-voxel connections followed a power-law scaling with an exponent close to 2, suggesting a scale-free network topology.
Instead, the analysis of networks under asymmetric link dynamics involves the link connectivity distribution, defined as the fraction of links connecting a pair of nodes with given connectivities.
Rather than resolving into several modules with a scale-free connectivity distribution, the vast majority of molecules are highly connected together into a single dense, cohesive network.
At the same time, this procedure preserves the number of connections that each node has, and thus the network's connectivity distribution.
This network also has a significant abundance of tetramer-like connection patterns, when compared to randomized networks with the same connectivity distribution.
Then, we drew 5,000 samples from the connectivity distribution for each voxel in the ROI and calculated the connection probability of each voxel.
