Table 1.

Tests of positive selection in CatSper1

N	Ls	Best fit model	Log-likelihood values a	Parameter estimates	PSS b
16	235	M1a	-2907.9808	p0 = 0.269, p1 = 0.731, ω0 = 0.027, ω1 = 1	Not allowed
		M2a	-2879.567332 **	p0 = 0.186, p1 = 0.651 p2 = 0.163, ω0 = 0, ω1 = 1, ω2 = 4.67	124H*, 128H*, 146S*, 147 T**, 154H**, 165S*, 179 L*, 198 T**, 201A**, 204Q**
		M7	-2908.013308	p = 0.041, q = 0.012	Not allowed
		M8	-2879.676624 **	p0 = 0.833, p = 0.0166, q =0.0052, p1 = 0.166, ω = 4.535	124H*, 146S*, 147 T**, 154H**, 165S*, 179 L*, 198 T**, 201A**, 204Q**
		M8a	-2907.982933	p0 = 0.269, p = 2.758, q = 99.0, p1 = 0.731, ω = 1.0	Not allowed

^aDifferences between log-likelihood values of models with 99% statistical significance level for 2 degrees of freedom are shown as **.

^b Positively selected sites (PSS) with a posterior probability >0.95 (*) and >0.99 (**) in Bayes Empirical Bayes.

N, Number of sequences aligned.

Ls, Sequence length (in codons) after alignment gaps are removed.

Parameter estimation and likelihood scores under site models.