TABLE 7

Simulation results summary

Marker:AFLPsSNPsMicrosatellites
Method used:IC (%)RJ (%)IC (%)RJ (%)IC (%)RJ (%)
Directional selection
False positive3.22.52.32.12.52.3
True positive70.870.465.667.394.194.0
PPV73.578.078.580.182.483.7
Balancing selection
False positive1.92.52.82.92.52.7
True positive33.740.245.849.678.781.2
PPV69.466.767.168.279.879.1
Total
False positive5.05.05.05.05.05.0
True positive52.355.355.758.586.487.6
PPV72.273.573.374.681.281.5
• A summary of the results for AFLP, SNP, and microsatellite data sets in Tables 35 is shown. The PPV is defined as the proportion of markers detected as being under selection that are correctly classified. Results are presented using cutoff values that lead to a 5% total false-positive rate for both the reversible-jump (RJ) method introduced here and the informal criterion (IC) originally proposed by Beaumont and Balding (2004).