Figure 5.

Receiver operating characteristic (ROC) analysis for the data sets 1 (A), 2 (B), 3 (C), 4 (D), 5 (E), 6 (F), 7 (G), 8 (H) and 9 (I). In the ROC analysis, the proportion of false positives and true positives is computed for each possible value of the threshold that is used to classify a locus under selection. For SelEstim, the classifying variable was the KLD between the posterior distribution of the locus-specific coefficient of selection δ_j and its centering distribution, while in the case of BayeScan it was the Bayes factor.