Figure 1

Receiver operating characteristic curves. Plot of the true-positive rate against the false-positive rate for the different possible cutpoints of a diagnostic test. These ROC curves are for the two family-based approaches and the population-based association analysis that we performed using GAW17 family-based data sets and population-based data sets. The shape of each curve indicates the quality of the corresponding method. The closer a ROC curve is to the diagonal of the plot, the worse the corresponding method is. A hypothetical perfect method would have a ROC curve that is a constant function with true-positive value 1 and false-positive value 0. The two family-based approaches (the logistic regression using GEE and the IBD analysis) performed significantly better than the standard population-based association study.