# Table 1 Parameter estimates under branch, site and branch-site models

Model p InL Estimate of parameters Models compared 2ΔL
Branch models
M0: ω0 = ω1 = ω2 = ωc 1 −10227.98 ω0 = ω1 = ω2 = ωc = 0.130
Mf: ω free 51 −10109.65 ω: 0 ~ 0.941 Mf vs. M0 236.66***
M2a: ω0 = ω2 = ωc, ω1 2 −10227.97 ω0 = ω2 = ωc =0.130, ω1 = 0.150 M2a vs. M0 0.02
M2b: ω01 = ωc, ω2 2 −10227.89 ω0 = ω1 = ωc = 0.129, ω2 = 0.190 M2b vs. M0 0.18
M2c: ω0 = ω1 = ω2, ωc 2 −10210.87 ω0 = ω1 = ω2 = 0.122, ωc = 0.951 M2c vs. M0 34.22***
M3a: ω0 = ωc, ω1, ω2 3 −10227.88 ω0 = ωc = 0.129, ω1 = 0.149, ω2 = 0.190 M3a vs. M2a 0.18
M3b: ω0 = ω2, ω1, ωc 3 −10210.85 ω0 = ω2 = 0.122, ω1 = 0.153, ω2 = 0.949 M3b vs. M2c 0.04
M3c: ω0 = ω1, ω2, ωc 3 −10212.06 ω0 = ω1 = 0.121, ω2 = 1.020, ωc = 0.756 M3c vs. M2c −2.38
M4a: ω0, ω1, ω2, ωc 4 −10210.82 ω0 = 0.122, ω1 = 0.153, ω2 = 0.148, ωc = 0.942 M4a vs. M2c 0.10
Site-specific models
M1: nearly neutral 1 −10152.15 P0 = 0.929, P1 = 0.071 ωO =0.109, ω1 = 1.000
M2: PositiveSelection 3 −10152.15 P0 = 0.929, P1 = 0.071, P2 = 0.000 ω0 = 0.109, ω1 = 1.000, ω2 =30.688 M2 vs. M1 0
M3: discrete (k=2) 5 −10064.21 P0 = 0.534, P1 = 0.417, P2 = 0.048 ωO = 0.028, ω1 = 0.223, ω2 = 0.670 M3 vs. M0 327.54***
M7: beta 2 −10064.80 P = 0.580 q = 3.405
M8: beta & ω 4 −10061.64 p0 = 0.989, p = 0.646, q = 4.203 (p1 = 0.011) ω = 1.054 M8 vs. M7 6.32*
Branch-site model
model A 3 −10130.51 P0 = 0.353, p1 = 0.026, p2 = 0.621, ω2 = 1.442 MA vs.M1 43.28***
1. * significant at p <0.05 level, *** significant at p <0.001 level.
2. p, the number of free parameters for the ω ratio. 