Figure 4

Line plot illustrating the effect of number of genome examples on the inductive CGP performance. This figure shows the effect of AUC versus number of genome examples in inductive CGP (SVM/Poly algorithm) on Streptococcus agalactiae 2603 genes related to peptidoglycan metabolism (M validation set). Twenty-five simulation runs of stratified 10-fold cross-validation were performed. The black line indicates the median AUC, the grey solid lines indicate the upper and lower quartiles, and the dotted grey lines indicate the maximum and minimum AUC of the simulation runs.