Table 2. Model comparison via log-likelihood ratio tests.
Null Model | Experimental Model | Probability | Testing for | Result |
Model 0 (one ratio) | Model 1a (nearly neutral) | <10−3 | Do s differ among sites? | Yes |
Model 1a (nearly neutral) | Model 2a (positive selection) | >0.999 | Is there positive selection? | No |
Model 0 (one ratio) | Model 3 (discrete) | <10−3 | Do s differ among sites? | Yes |
Model 7 (beta) | Model 8 (beta & ) | 0.004 | Is there positive selection? | Yes |