Table 1 PAML results: log likelihoods, parameter estimates, Bayes Empirical Bayes identified sites, and likelihood ratio tests.

N/A

M0 (one-ratio)

33

-2127.37

ω = 0.08737

none

N/A

Site-specific models

M1a (neutral)

34

-2097.31

p₀ = 0.91742, ω₀ = 0.05054

N/A

N/A

M2a (selection)

36

-2097.31

p₀ = 0.91768, ω₀ = 0.05052, p₂ = 0.04090, ω₂ = 0.99980

275

P = 1 (vs. m1a)

M3 (discrete, K = 3)

37

-2088.38

p₀ = 0.53422, p₁ = 0.38154, ω₀ = 0.00000, ω₁ = 0.12511, ω₂ = 0.63618

none

P < 0.000001 (vs. m0)

M7 (β)

34

-2089.48

p = 0.21896, q = 1.82556

N/A

N/A

M8 (β & ω)

36

-2088.91

p = 0.24125, q = 2.23372, p₀ = 0.99249, ω = 1.66242

275

P = 0.568523 (vs. m7)

Branch-specific models

X 'green'

two-ratio

34

-2123.10

ω₀ = 0.0806, ω₁ = 0.4259

N/A

P = 0.003463 (vs. m0)

Y 'transducin'

two-ratio

34

-2125.68

ω₀ = 0.0832, ω₁ = 0.4523

N/A

P = 0.066388 (vs. m0)

Branch-site (BS) models

X 'green'

BS model A (null)

35

-2094.55

p₀ = 0.73148, ω₀ = 0.04521, p₁ = 0.06817

N/A

N/A

BS model A (alternative)

36

-2093.93

p₀ = 0.86007, ω₀ = 0.04576, p₁ = 0.06817, ω₂ = 16.53309

271, 275, 277, 279, 285, 286, 302

P = 0.266644 (vs. BS model A [null])

Y 'transducin'

BS model A (null)

35

-2093.10

p₀ = 0.31761, ω₀ = 0.04432, p₁ = 0.03013

N/A

N/A

BS model A (alternative)

36

-2093.10

p₀ = 0.32723, ω₀ = 0.04442, p₁ = 0.03012, ω₂ = 1.00000

248, 249, 256, 261, 263

P = 1 (vs. BS model A [null])