TABLE 6.

Simulation results for the method without Bayesian variable selection

	True:	Balancing selection			Neutrality			Directional selection
	Classified:	Balancing	Neutrality	Directional	Balancing	Neutrality	Directional	Balancing	Neutrality	Directional	Specificity	Power
s100		10	38	2	57	840	3	0	0	50	0.93	0.60
s050		12	38	0	62	830	8	0	0	50	0.92	0.62
s020		10	38	2	37	856	7	0	0	50	0.95	0.60
s100-Fb		13	36	1	66	824	10	0	0	50	0.92	0.63
s050-Fb		11	38	1	68	829	3	0	0	50	0.92	0.61
s020-Fb		3	46	1	83	806	11	0	0	50	0.90	0.53
s100-Fb-40		6	43	1	71	823	6	0	0	50	0.91	0.56
Neutral					20	941	39				0.94

Numbers of loci simulated under balancing selection, neutrality, and directional selection that were classified in each category by the reparameterized Bayesian regression analysis without variable selection are shown. A locus was classified as “directional” if P(α_i < 0 | data) ≤ 0.05, as “balancing” if P(α_i < 0 | data) ≥ 0.95, and otherwise as “neutral.”