Model | No. of parameters | Estimates of parameters^{a} | ln L | 2Δ(ln L)^{c} | P-value |
---|---|---|---|---|---|

M0: one ratio | 1 | ω_{0} = 0.04 | −2393.88 | ||

M1a: neutral | 1 | p_{0} = 0.94, p_{1} = 0.06 | −2353.24 | ||

ω_{0} = 0.02, ω_{1} = 1.00 | |||||

M2a: positive selection | 3 | p_{0} = 0.93, p_{1} = 0.06, p_{2} = 0.01 | −2351.08 | 4.31 | 0.116 |

ω_{0} = 0.02, ω_{1} = 1.00, ω_{2} = 4.70 | |||||

M3: discrete | 5 | p_{0} = 0.84, p_{1} = 0.15, p_{2} = 0.01 | −2335.11 | ||

ω_{0} = 0.01, ω_{1} = 0.25, ω_{2} = 3.38 | |||||

M7: β | 2 | ω from β(0.10, 1.44)^{b} | −2341.55 | ||

M8: β and ω | 4 | p_{0} = 0.99, ω from β(0.11, 2.07) | −2334.94 | 13.22 | 0.0014 |

(p_{1} = 0.01) ω_{1} = 3.60 |

↵

*a*The proportion of sites (p_{0}, p_{1},…) estimated to have ω_{0}, ω_{1}, … .↵

*b*In M7 and M8 10 ω-values that are <1 are estimated from a β-distribution defined by β(*x*,*y*).↵

*c*Significance was assessed as two times the difference in log likelihoods of corresponding positive selection and neutral models.