Table 2.

Power calculation under Scenario (B)

Strategy	Corr	MAF	Method
			LR	FPM	TPM	Lasso	Enet	GRLR
I	0.0	0.10	0.302	0.346	0.320	0.330	0.388	0.310
		0.30	0.740	0.716	0.732	0.826	0.832	0.696
		0.50	0.906	0.888	0.900	0.932	0.940	0.872
	0.2	0.10	0.342	0.422	0.422	0.436	0.404	0.368
		0.30	0.814	0.846	0.832	0.838	0.854	0.784
		0.50	0.948	0.942	0.946	0.960	0.966	0.900
II	0.0	0.10	0.040	0.200	0.180	0.158	0.202	0.270
		0.30	0.280	0.478	0.490	0.630	0.704	0.680
		0.50	0.544	0.622	0.612	0.804	0.862	0.856
	0.2	0.10	0.058	0.256	0.264	0.210	0.238	0.314
		0.30	0.440	0.580	0.582	0.646	0.730	0.758
		0.50	0.700	0.672	0.678	0.858	0.914	0.890
III	0.0	0.10	0.002	0.022	0.016	0.044	0.066	0.100
		0.30	0.030	0.082	0.066	0.340	0.434	0.460
		0.50	0.140	0.084	0.076	0.558	0.640	0.656
	0.2	0.10	0.004	0.056	0.050	0.062	0.080	0.122
		0.30	0.124	0.088	0.090	0.324	0.422	0.490
		0.50	0.292	0.094	0.092	0.550	0.672	0.704

Results of power simulations for six methods under Scenario (B) (three causal SNPs). Three definitions of testing success, Strategy I (requiring that at least one of the three SNPs is significant), Strategy II (requiring that at least two of the three SNPs are significant), and Strategy III (requiring significance at all three SNPs), are applied in comparisons; The number of replication runs is 500 for all the simulations. Corr = correlation coefficient; MAF = minor allele frequency.