Table 4.

Empirical type I error rates (%) and power (%) under four discrete populations at 1% significance level

	Configuration	m	Trend	ES	PS	MC	SS
Non-causal SNPs: random (category 1)	C5	100	46.22	1.3	1.09	1.22	0.53
		300	46.6	1.04	1.06	1.04	0.41
		500	46.17	1.02	1.05	1.02	0.22
	C6	100	64.21	3.31	1.19	3.52	2.19
		300	64.09	1.18	1.06	1.42	0.87
		500	64.11	1.11	1.05	1.15	0.75
Non-causal SNPs: highly differentiated (category 2)	C5	100	34.56	1.14	1.07	1.16	0.97
		300	34.41	0.96	1.04	1.03	0.93
		500	34.39	1.02	0.98	1.04	0.92
	C6	100	100	13.45	1.21	10.19	6.67
		300	100	2.2	1.09	2.06	0.68
		500	100	1.55	1.11	1.45	0.39
Causal SNPs: random (category 3)	C5	100	64.42	86.82	86.42	85.67	74.42
		300	64.24	86.06	86.33	86.28	74.73
		500	64.02	86.34	86.46	86.49	74.91
	C6	100	70.54	70.14	71.86	71.11	62.16
		300	70.33	71.58	71.98	72.06	63.86
		500	70.18	71.4	71.9	72.26	64.2
Causal SNPs: highly differentiated (category 4)	C5	100	0.04	73.09	77.64	77.77	35.05
		300	0.05	76.32	77.91	79.37	37.4
		500	0.05	78.7	78.6	79.88	38.59
	C6	100	100	11.52	40.03	25.55	4.61
		300	100	39.06	51.27	39.49	20.77
		500	100	46.78	52.45	52.18	28.08

m is the number of random SNPs used for inference of population structure. ES: EIGENSTRAT; PS: PHYLOSTRAT; MC: MDS clustering; SS: STRATSCORE.