Table 3.

Percentage of datasets with the highest scoring marker within distance ∊ (in kb) from the disease-predisposing SNP, which is untyped

∊	GRR=1.4			GRR=1.6			GRR=1.8			GRR=2.0
	χ2	P	U	χ2	P	U	χ2	P	U	χ2	P	U
500 case and 500 control individuals
10	38	23	19	44	41	33	41	47	45	42	51	45
20	50	47	42	64	62	57	61	70	64	62	72	69
30	59	64	54	77	69	68	69	82	79	73	82	84
40	69	71	69	82	79	80	77	86	86	80	88	91
50	78	81	82	89	85	85	81	89	91	85	91	93
60	85	86	86	90	88	89	88	94	93	94	94	97
70	92	92	90	94	92	94	93	95	94	97	96	98
80	98	96	97	98	97	99	98	97	97	99	97	99
2000 case and 2000 control individuals
10	44	48	45	44	52	52	49	65	68	57	72	74
20	68	63	66	62	72	69	72	78	79	77	88	88
30	76	81	80	74	88	82	79	88	88	89	95	94
40	84	87	90	84	92	87	86	92	89	95	96	97
50	88	93	97	94	96	91	91	94	92	96	98	99
60	95	97	98	97	99	96	95	98	97	97	99	100

Based on 100 simulated datasets for each setting, with 5% WR and ρ=40. Columns denoted χ² correspond to single-marker χ²-test results, while columns denoted ‘P’ (‘U’, respectively) correspond to Blossoc results using m=5 for phased (unphased, respectively) data.