Table 1

Comparison of alternatively parameterized models of evolution in Fig. 1 indicates no significant difference in the Ka/Ks ratio at an early and a later stage of duplicate gene evolution.

Comparison

# base pairs

-lnL Ho

-lnL Ha

P value

Ka/Ks combined early and late

Ka/Ks ratio early

Ka/Ks ratio late

Ka/Ks diploid


Fig. 1B

80856

-165602.720

-165602.386

0.414

0.164

0.158

0.169

0.126

Fig. 1C

9717

-15699.366

-15697.250

1.000

0.208

0.124

0.346

0.198

Fig. 1D

6966

-13187.865

-13186.872

0.160

0.126

0.187

0.105

NA

Fig. 1B (partitioned)

80856

-160085.863

-159889.926

1.000

NL

Af2

Af2

NL

Fig. 1C (partitioned)

9717

-15400.349

-15393.089

0.888

NL

Af2

Af2

NL

Fig. 1D (partitioned)

6966

-12983.343

-12978.034

0.807

NL

Af2

Af2

NA


Indicated for comparisons depicted in Fig. 1B, C and D are likelihoods of the null model (early and later Ka/Ks are the same) and the alternative model (early and later Ka/Ks are not the same), the one-sided probability of the Ka/Ks ratio being higher in the early stage, and the Ka/Ks ratios estimated from each of these models. For the first two tests, the Ka/Ks ratio of the diploid lineage was estimated using a different model where a unique Ka/Ks ratio was estimated for each branch (a free ratio model). Also listed are the joint likelihoods of these models from an analysis partitioned by gene fragment. For the partitioned analyses, Ka/Ks ratios for each fragment are either listed in Additional file 2 (Af2), not listed (NL), or not applicable (NA).

Chain et al. BMC Evolutionary Biology 2008 8:43   doi:10.1186/1471-2148-8-43

Open Data