A non-parametric approach for detecting gene-gene interactions associated with age-at-onset outcomes

BMC Genomic Data

Table 2 Comparison of WNA and COX in prospective cohort studies

Disease models							Two Causal SNP		+ 2 Noise SNPs		+5 Noise SNPs		+ 8 Noise SNPs
Disease models							WNA	COX	WNA	COX	WNA	COX	WNA	COX
Model 1: PH; Linear	θ = 3
	β₁ = 0.15					Power:	0.702	0.738	0.641	0.660	0.596	0.532	0.567	0.427
	β₂ = 0.15					Sensitivity:	0.736	0.924	0.686	0.926	0.633	0.906	0.587	0.906
	β₁₂ = 0.15					Specificity	--	--	0.967	0.966	0.942	0.904	0.932	0.847
Model 2: PH; non-linear						Power:	0.802	0.623	0.731	0.513	0.687	0.386	0.661	0.297
	θ = 3					Sensitivity:	0.781	0.818	0.730	0.810	0.681	0.810	0.647	0.810
	β = 0.6					Specificity	--	--	0.975	0.970	0.956	0.905	0.940	0.869
Model 3: Non-PH	θ		AA	Aa	aa
		BB	3	2.5	2.5	Power:	0.932	0.733	0.910	0.652	0.903	0.522	0.891	0.428
		Bb	3	2	2	Sensitivity:	0.611	0.778	0.588	0.786	0.573	0.786	0.556	0.786
		bb	3	2	2	Specificity	--	--	0.987	0.963	0.974	0.903	0.965	0.843
Model 4: Non-PH	θ		AA	Aa	aa
		BB	3	3	1	Power:	0.989	0.508	0.979	0.420	0.971	0.304	0.958	0.232
		Bb	3	3	1	Sensitivity:	0.916	0.769	0.851	0.773	0.779	0.773	0.737	0.769
		bb	1	1	0.5	Specificity	--	--	0.897	0.860	0.836	0.799	0.802	0.739

ISSN: 2730-6844