Table 3.
Test | Line | Alternative Ma | Null Ma | 2ΔL | df | Pb | Sitesc |
---|---|---|---|---|---|---|---|
1 | a | Nearly Neutral | One ratio | 13.75 | 1 | 0.0002 | 4 |
2 | b | Local relaxation | Uniform Selection | 5.39 | 2 | 0.02 | 0 |
3-D | c | Local relaxation | Uniform Selection | 64.89 | 2 | < 0.0001 | 16 |
3-T | d | Local relaxation | Uniform Selection | 15.79 | 2 | < 0.0001 | 1 |
a Alternative and null models, see Additional File 6, Table S3 for more information on models and Log-likelihood values, b P value calculated from a χ2 distribution, c number of sites significant in Bayesian post-hoc tests (P < 0.05). The alignment, after deleting gaps, was composed of 122 codons. Tests that we deemed weakly significant because Bayesian post-hoc tests did not detect relevant AA are shown in italics.