Table 1.
Likelihood ratio test statistics for models of variable selective pressure among sites (F61 model of codon frequency).
| Region/selection model(number of codons) | d.f. | -2ΔLnL | P value | % of sites(average dN/dS) | Selected sites |
|---|---|---|---|---|---|
| Loop1 (269) | |||||
| M7 versus M8 | 2 | 10.04 | 0.0066 | 1.3% (1.37) | 182 (0.99, n.s.) |
| M8a versus M8 | 1 | 4.87 | 0.027 | - | - |
| Loop 2 N-term (251) | |||||
| M7 versus M8 | 2 | 25.86 | <0.0001 | 1.3% (1.78) | 416 (0.99, 0.99),417 (0.98, 1), 421 (0.91, 0.99) |
| M8a versus M8 | 1 | 10.98 | 0.0009 | - | |
| Loop 2 C-term (81) | |||||
| M7 versus M8 | 2 | <0.01 | >0.99 | - | - |
| M8a versus M8 | 1 | 1.96 | 0.16 | - | |
| SSD (164) | |||||
| M7 versus M8 | 2 | 5.70 | 0.058 | - | - |
| M8a versus M8 | 1 | 0.02 | >0.88 | - | - |
| Loop 3 (248) | |||||
| M7 versus M8 | 2 | 0.004 | >0.99 | - | - |
| M8a versus M8 | 1 | 1.52 | 0.21 | - | - |
M7 is a null model that assumes that 0<ω<1 is beta distributed among sites; M8a assumes 0<ω≤1; these two models are compared to M8, which includes an extra category of sites with ω>1 (positive selection). d.f., degree of freedom; 2ΔLnL, twice the difference of the natural logs of the maximum likelihood of the models being compared; P value, P value of rejecting the neutral models (M7 or M8a) in favor of the positive selection model (M8); % of sites (average dN/dS), estimated percentage of sites evolving under positive selection by M8 (dN/dS for these codons); selected sites, the position refers to the entire NPC1 human sequence, P values obtained from BEB and REL analyses are shown. BEB, Bayes empirical Bayes; REL, random effects likelihood; SSD, sterol sensing domain.