Figure 16

A: Network learned by Deterministic Effects Propagation Networks purely from data. Gray nodes indicate latent nodes without measurements. Dashed arrows were inferred from data, but cannot be explained by current literature knowledge. For the sake of better visualization we omit indirect edges between nodes, i.e. interactions, which can be explained by a cascade of others (transitive reduction). Numbers at edges indicate bootstrap probabilities. B: Network learned by Deterministic Effects Propagation Networks from data when inference started with known interactions. The dashed line indicates an interaction that cannot be explained by the literature network. Numbers at edges indicate bootstrap probabilities.