Skip to main content

Table 3 Comparison of WNA and MCC in case-control studies

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

Disease Models

 

Two Causal SNP

+ 2 Noise SNPs

+5 Noise SNPs

+ 8 Noise SNPs

WNA

MCC

WNA

MCC

WNA

MCC

WNA

MCC

Model 1: PH; Linear

θ = 3

         

β1 = 0.15

Power:

0.762

0.614

0.718

0.434

0.692

0.222

0.674

0.100

β2 = 0.15

Sensitivity:

0.753

0.980

0.717

0.980

0.693

0.978

0.680

0.978

β12 = 0.15

Specificity

--

--

0.926

0.913

0.827

0.763

0.753

0.627

Model 2: PH; non-linear

θ = 3

Power:

0.876

0.561

0.823

0.376

0.838

0.195

0.821

0.076

Sensitivity:

0.916

0.934

0.909

0.902

0.873

0.902

0.844

0.902

β = 0.6

Specificity

--

--

0.879

0.878

0.736

0.735

0.605

0.578

Model 3: Non-PH

θ

 

AA

Aa

aa

         

BB

3

2.5

2.5

Power:

0.926

0.678

0.933

0.480

0.930

0.253

0.932

0.114

Bb

3

2

2

Sensitivity:

0.753

0.891

0.717

0.891

0.693

0.891

0.680

0.891

bb

3

2

2

Specificity

--

--

0.926

0.915

0.827

0.777

0.753

0.629

Model 4: Non-PH

θ

 

AA

Aa

aa

         

BB

3

3

1

Power:

0.987

0.458

0.974

0.277

0.960

0.133

0.952

0.057

Bb

3

3

1

Sensitivity:

0.971

0.831

0.905

0.831

0.831

0.831

0.769

0.830

  

bb

1

1

0.5

Specificity

--

--

0.878

0.917

0.781

0.790

0.729

0.656