Fig. 2.

Evaluation of PPI inference for two AP-MS representation models. A, schematic representation of two possible models, spoke and matrix, used to annotate PPIs from AP-MS experiments. B, precision/recall curve as measured from the cross-validation procedure. True positives (TP) (correctly inferred gold standard positive interactions), false positives (incorrectly inferred as interacting gold standard negative interactions), and FN (not inferred as interacting gold standard positive interactions). C, precision as the function of the number of PPIs inferred by the logistic regression classifier.