Figure 5

Study 5 undirected edge posterior probabilities. For the fifth study, (a), (b), (c) and (d) are representations of the undirected edge posterior probabilities. These are computed for the composite of all the replicates, C⁽⁵⁾, and for the individual replicates, $R_1^{\left(5\right)}$, $R_2^{\left(5\right)}$ and $R_3^{\left(5\right)}$. Blocks of vertices $\left\{p_1^{\left(5\right)},p_2^{\left(5\right)},p_3^{\left(5\right)}\right\}$, $\left\{p_4^{\left(5\right)},p_5^{\left(5\right)},p_6^{\left(5\right)}\right\}$, $\left\{p_7^{\left(5\right)},p_8^{\left(5\right)},p_9^{\left(5\right)}\right\}$ and $\left\{p_{10}^{\left(5\right)},p_{11}^{\left(5\right)},p_{12}^{\left(5\right)}\right\}$ are signal correlated in $R_1^{\left(5\right)}$ are with ρ = 0.9, in $R_2^{\left(5\right)}$ with ρ = 0.82, and in $R_3^{\left(5\right)}$ with ρ = 0.7 (3). In the diagrams proteins $p_1^{\left(5\right)},\dots ,p_{12}^{\left(5\right)}$ are represented by A, . . . , L for ease of reading. The composite has high posterior probabilities for all edges with at least moderate network partial correlations. The composite model outperforms each of the individual replicates' models.