TABLE 2.

Effect of the number of independent loci on the estimation of the admixture rate λ by the ABC and Wang's (2003) methods

		ABC
100,000 simulations				1,000,000 simulations				WANG03
Sample size	No. of loci	Bias	RMSE	Coverage 95%	Factor 2	Bias	RMSE	Coverage 95%	Factor 2	Bias	RMSE	Coverage 95%	Factor 2
20	5	−0.057	0.232	0.99	0.98	−0.029	0.228	0.99	0.99	0.003	0.212	0.95	0.98
10	−0.055	0.181	1.00	0.99	−0.046	0.171	1.00	1.00	0.016	0.157	0.89	1.00
20	−0.044	0.125	1.00	1.00	−0.054	0.118	0.99	1.00	−0.010	0.114	0.93	1.00
50	−0.012	0.071	1.00	1.00	−0.010	0.074	0.99	1.00	0.003	0.074	0.93	1.00
100	5	−0.089	0.175	1.00	1.00	−0.078	0.162	1.00	1.00	−0.026	0.134	0.89	1.00
10	−0.051	0.125	1.00	1.00	−0.043	0.112	1.00	1.00	−0.009	0.091	0.86	1.00
20	−0.035	0.091	1.00	1.00	−0.025	0.079	0.99	1.00	−0.009	0.070	0.87	1.00
	50	−0.003	0.058	1.00	1.00	−0.010	0.042	0.99	1.00	−0.004	0.039	0.89	1.00

Simulated conditions are independent loci. λ = 0.3, t_ADM = 5, t_DIV = 5000, N₀ = N₁ = N₂ = 300. Bias and root mean square error (RMSE) are expressed in relative units. Coverage 95% represents the number of times among 100 that the true value of λ (0.3) lies within the estimated 95% confidence interval. Factor 2 represents the number of times that the true value of λ lies within an interval limited by 50 and 200% of the estimated λ value.