Table 3. Patterns of diversity at selected sites in the focal region.

γ	r/u	Model	π			θW			Tajima's D			Fu and Li's D
			Mean	s.d. ( × 10−4)	Cor	Mean	s.d. ( × 10−4)	Cor	Mean	s.d.	Cor	Mean	s.d.	Cor
15	0	Forw	6.7 × 10−4	2.88	0.028	0.0015	5.88	0.082	−1.741	0.342	0.083	−2.312	0.918	0.144
		Coal	6.5 × 10−4	2.12	0.021	0.0015	4.96	0.064	−1.793	0.324	0.079	−2.417	0.910	0.157
15	3	Forw	6.7 × 10−4	2.81	0.026	0.0015	5.79	0.071	−1.755	0.329	0.072	−2.336	0.896	0.112
		Coal	6.5 × 10−4	2.12	0.019	0.0015	4.88	0.057	−1.811	0.305	0.076	−2.449	0.883	0.114
7.5	0	Forw	0.0014	5.44	0.242	0.0025	8.21	0.233	−1.542	0.430	0.274	−1.977	0.971	0.279
		Coal	0.0013	2.91	0.063	0.0025	6.47	0.139	−1.712	0.341	0.191	−2.228	0.922	0.288
7.5	3	Forw	0.0014	4.67	0.084	0.0025	7.70	0.126	−1.625	0.367	0.115	−2.064	0.901	0.158
		Coal	0.0013	2.91	0.031	0.0026	6.30	0.088	−1.777	0.306	0.108	−2.311	0.863	0.159

In all cases, the focal region is composed of 2001 sites. Every third site is neutrally evolving, and all the other sites are subject to selection with the same γ. The scaled mutation is θ=0.005 per site and the sample size is 20. The mean and s.d. of the four statistics were estimated using sequence variability at the selected sites alone. In contrast, the correlation coefficient (Cor) was estimated using data from both the selected and neutral sites. Note that the expected value of π in an infinite population is 6.7 × 10⁻⁴ for γ=15 and 0.0013 for γ=7.5.