TABLE 3

Performance of Adaptsite and codeml in inferring positive selection sites in simulated data

5 taxa (tree A)30 taxa (tree B)
Simulation scheme         TestProportion of
 true positivesProportion of
 false positivesProportion of
 true positivesProportion of
 false positives
Scheme 4Adaptsite0.000.000.000.00
 (100 replicates,Codeml (before LRT)M20.080.080.270.24
 50% ω = 1.5,M2a0.450.420.320.28
 50% ω = 1)M80.380.360.190.16
Codeml (after LRT)M1 vs. M20.070.060.230.21
M1a vs. M2a0.340.320.290.25
M7 vs. M80.280.260.160.14
M8a vs. M80.350.330.190.16
M8a vs. M8 (1 d.f.)0.350.330.190.16
Scheme 5Adaptsite0.000.000.000.00
 (50 replicates,Codeml (before LRT)M20.410.200.200.10
 45% ω = 0,M2a0.440.210.300.14
 45% ω = 1,M80.130.050.090.03
 10% ω = 1.5)Codeml (after LRT)M1 vs. M20.000.000.020.01
M1a vs. M2a0.040.020.040.02
M7 vs. M80.100.040.090.03
M8a vs. M80.050.040.060.01
M8a vs. M8 (1 d.f.)0.050.040.060.01
Scheme 6Adaptsite0.000.000.080.00
 (50 replicates,Codeml (before LRT)M20.190.000.760.00
 45% ω = 0,M2a0.180.000.750.00
 45% ω = 1,M80.200.000.760.00
 10% ω = 5)Codeml (after LRT)M1 vs. M20.190.000.760.00
M1a vs. M2a0.180.000.750.00
M7 vs. M80.200.000.760.00
M8a vs. M80.200.000.760.00
M8a vs. M8 (1 d.f.)0.200.000.760.00
  • The proportion of true positives is defined as the number of sites that are correctly classified as positively selected divided by the total number of positive selection sites simulated. The proportion of false positives is defined as the number of sites that are falsely classified in the positively selected category divided by the total number of sites that are not positively selected.