The results of our full Scenario B comparison are shown in Figure 5 in two panels that mirror Figure 4A and 4B. Although this dataset contains a large number of HapMap-only SNPs that were imputed here but not in the restricted dataset, we calculated discordance only at masked chip SNPs in the study sample, so the curves in Figure 4 and Figure 5 are based on exactly the same sets of masked genotypes. There are four curves in each panel: IMPUTE v2 in the full Scenario B dataset (k = 40; dashed red line); BEAGLE in the full dataset (default settings; dashed green line); BEAGLE in the restricted dataset (default settings; solid green line); and IMPUTE v1 in the restricted dataset (solid blue line). The first two curves (dashed lines) are the main focus of this comparison, and the latter two curves (solid lines) are carried over from Figure 4 for reference.
