Table 5. Likelihood ratio statistics of different models.

Locus	Comparison	Df	lnL	X2 Critical value (1%)
GP43	One-ratio (M0) vs. Discrete (M 3)	4	10.966608	9.21
	Neutral (M 1) vs. Selection (M 2)	2	18.10675	13.28
	Beta (M 7) vs. Beta+w (M 8)	2	18.641266	9.21
FKS	One-ratio (M0) vs. Discrete (M 3)	4	0	9.21
	Neutral (M 1) vs. Selection (M 2)	2	0	13.28
	Beta (M 7) vs. Beta+w (M 8)	2	0	9.21
CDC42	One-ratio (M0) vs. Discrete (M 3)	4	1.4568	9.21
	Neutral (M 1) vs. Selection (M 2)	2	2.676	13.28
	Beta (M 7) vs. Beta+w (M 8)	2	2.6354	9.21
P27	One-ratio (M0) vs. Discrete (M 3)	4	0	9.21
	Neutral (M 1) vs. Selection (M 2)	2	0	13.28
	Beta (M 7) vs. Beta+w (M 8)	2	0	9.21

Twice the difference in log likelihood ratio between a null model and an alternative model was compared with a χ² distribution in order to test whether an alternative model fits the data better than the null model. Df: Degrees of Freedom; LRT: Likelihood Ratio Test.