Figure 2

Yeast KID performance in identifying kinase-substrate pairs. (a) The graph indicates the true positive rate detected by the number of experimental evidence supporting a kinase-substrate interaction reported in KID and BioGRID [21, 22] and the top scores reported in both databases for kinases, as a function of their false positive rates (ROC curve). The diagonal line represents the random assignment of positive classes. The green line shows the cutoff score used for the stringent gold standard of kinase-substrate pairs. (b) The precision of the number of experimental evidence supporting kinase-substrate interactions and the respective top scores reported in KID and BioGRID [21, 22] for yeast kinases, as a function of their recall (equivalent to the true positive rate). The performance of a random assignment of positive classes is not shown as it is too low for representation.