Skip to main content
. Author manuscript; available in PMC: 2012 Oct 9.
Published in final edited form as: Stat Interface. 2011;4(3):359–371. doi: 10.4310/sii.2011.v4.n3.a10

Table 6. The median type I error probabilities and median observed powers of the FNR-based approach in stage one analysis for the five simulation models, given pre-specified multi-testing powers (50%, 60%, 70%, 80%and 90%) and different parameter λ values (10-5, 5×10-5, 10-4, and 10-3), based on 100 replicates.

Pre-specified multi-testing powers
50% 60% 70% 80% 90%
Model
(# of
causal
SNPs)		 Parameter
values		 Type I
errors		 Observed
powers		 Type I
errors		 Observed
powers		 Type I
errors		 Observed
powers		 Type I
errors		 Observed
Powers		 Type I
errors		 Observed
powers
Model 1 (5) λ = 10−5 0 0.60 2.00E-06 0.60 4.00E-06 0.60 6.00E-06 0.60 1.00E-05 0.60
λ = 5 × 10−5 6.00E-06 0.60 8.00E-06 0.60 1.20E-05 0.60 1.60E-05 0.60 2.20E-05 0.60
λ = 10−4 1.50E-05 0.60 2.40E-05 0.60 3.50E-05 0.60 4.90E-05 0.60 6.20E-05 0.60
λ = 10−3 2.51E-04 0.80 3.51E-04 0.80 4.37E-04 0.80 5.61E-04 0.80 6.60E-04 0.80
Model 2 (10) λ = 10−5 4.00E-06 0.50 6.00E-06 0.50 8.00E-06 0.50 1.20E-05 0.60 1.60E-05 0.60
λ = 5 × 10−5 1.20E-05 0.50 1.90E-05 0.60 3.20E-05 0.60 4.20E-05 0.60 5.60E-05 0.60
λ = 10−4 2.50E-05 0.60 3.60E-05 0.60 5.50E-05 0.70 7.80E-05 0.70 9.80E-05 0.70
λ = 10−3 3.31E-04 0.70 4.76E-04 0.80 6.11E-04 0.80 7.75E-04 0.80 9.38E-04 0.80
Model 3 (15) λ = 10−5 8.00E-06 0.53 1.20E-05 0.60 1.80E-05 0.60 2.40E-05 0.60 3.10E-05 0.60
λ = 5 × 10−5 2.80E-05 0.60 4.20E-05 0.67 5.70E-05 0.67 7.10E-05 0.67 9.10E-05 0.67
λ = 10−4 5.10E-05 0.67 7.10E-05 0.67 9.50E-05 0.67 1.20E-04 0.67 1.49E-04 0.67
λ = 10−3 3.65E-04 0.73 5.21E-04 0.73 6.77E-04 0.73 8.40E-04 0.73 1.05E-03 0.73
Model 4 (20) λ = 10−5 1.70E-05 0.55 2.40E-05 0.55 3.30E-05 0.55 4.20E-05 0.60 5.20E-05 0.60
λ = 5 × 10−5 4.50E-05 0.60 6.20E-05 0.60 7.90E-05 0.65 1.00E-04 0.65 1.21E-04 0.65
λ = 10−4 7.20E-05 0.60 9.70E-05 0.65 1.25E-04 0.65 1.56E-04 0.65 1.95E-04 0.70
λ = 10−3 3.55E-04 0.70 5.12E-04 0.75 6.82E-04 0.75 8.82E-04 0.75 1.11E-03 0.75
Model 5 (25) λ = 10−5 8.00E-06 0.40 1.30E-05 0.44 1.80E-05 0.48 2.40E-05 0.48 3.20E-05 0.52
λ = 5 × 10−5 3.10E-05 0.52 4.40E-05 0.56 5.80E-05 0.56 7.50E-05 0.60 9.30E-05 0.60
λ = 10−4 5.40E-05 0.56 7.40E-05 0.56 9.90E-05 0.60 1.30E-04 0.60 1.59E-04 0.64
λ = 10−3 3.91E-04 0.68 5.42E-04 0.68 7.22E-04 0.72 8.86E-04 0.72 1.10E-03 0.72

λ: parameter used for estimating number of null hypotheses.