Overview of analyses on network inference.

<p>Overview of all results on the simulated and biological datasets, using the approach presented in this manuscript. See the main text for comparison with other methods. Shown are results for the full network reconstruction task; <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0035077#pone-0035077-t003" target="_blank">table 3</a> shows corresponding results for hub identification. Each row in the table corresponds to one dataset. Nodes, edges and TP gives the number of genes, regulations and time points in the respective dataset. ROC and PR are the area under the curve values (AUC) of the Receiver Operator Characteristic (ROC) and Precision-Recall (PR) analysis, respectively. P-values were computed to test the null hypothesis of a significant deviation from random guessing for the AUC ROC values. Due to runtime limitations, MCMC results were calculated only for small networks, and p-values only for the synthetic networks with AUC values0.5. unk.: True number of edges for Yeast CC Network is unknown.