Table 2. Log-likelihood values (lnL), parameter estimates and likelihood ratio tests for various Branch-based models testing for heterogeneous selection pressures for melanopsin and rhodopsin.
Branch model | ω0 | ωPa | ωGb | lnLa | npb | H0 | H1 | 2ΔlnLc | d.f. | P-value |
---|---|---|---|---|---|---|---|---|---|---|
Melanopsin (opn4m2) | ||||||||||
A. One ratio: all branches have the same ω | 0.08 | ω0 | ω0 | −5800.498 | 17 | |||||
B. One ratio: all branches have the same ω=1 | 1 | ω0 | ω0 | −6407.852 | 16 | B | A | 1214.708 | 1 | 3.88E−266 |
C. Two ratio: (ω0), ωPa | 0.077 | 1.189 | ω0 | −5794.580 | 18 | A | C | 11.836 | 1 | 0.0006 |
D. Two ratio: (ω0), ωPa =1 | 0.077 | 1 | ω0 | −5794.597 | 17 | D | C | 0.033 | 1 | 0.856 |
E. Two ratio: (ω0), ωGb | 0.079 | ω0 | 0.167 | −5800.181 | 18 | A | E | 0.634 | 1 | 0.426 |
F. Free ratio (each branch has one ω) | Variable by branchd | 1.127 | 0.137 | −5777.164 | 31 | C | F | 34.832 | 13 | 0.001 |
Rhodopsin (rho) | ||||||||||
A. One ratio: all branches have the same ω | 0.055 | ω0 | ω0 | −4379.978 | 17 | |||||
B. One ratio: all branches have the same ω=1 | 1 | ω0 | ω0 | −4971.150 | 16 | B | A | 1182.344 | 1 | 4.19E−259 |
C. Two ratio (ω0), ωPa | 0.053 | 0.209 | ω0 | −4378.015 | 18 | A | C | 3.926 | 1 | 0.048 |
D. Two ratio: (ω0), ωPa =1 | 0.053 | 1 | ω0 | −4379.183 | 17 | D | C | 2.337 | 1 | 0.126 |
E. Two ratio: (ω0), ωGb | 0.054 | ω0 | 0.110 | −4379.770 | 18 | A | E | 0.414 | 1 | 0.520 |
F. Free ratio (each branch has one ω) | Variable by branchd | 0.199 | 0.072 | −4365.531 | 31 | C | F | 24.968 | 13 | 0.023 |
Abbreviations: Gb, Garra barreimiae; H0 and H1, null and alternative hypothesis; Pa, Phreatichthys andruzzii.
The best-fit model is underlined.
ln(likelihood) value.
Number of parameters.
Twice the difference of ln(likelihood) between the two models compared.
(N*dN) and (S*dS) values are shown in Figures 1a and b.