Figure 1.

CNs and quadrangular closure in bipartite networks. (A) A bipartite network of targets (t, dark green squares) and drugs (d, red circles) and its equivalent matrix representation. (B) Four CNs (white nodes) together with three LCLs (white links) as defined by Daminelli et al. [30] are used to calculate the interaction likelihood of the two seed nodes d_z and t_y (black nodes) of the existing link d_z-t_y (left). Similarly, five CNs and five LCLs are used to predict the likelihood of the missing d_x-t_y interaction (right) .