Skip to main content

Table 1 Degree of Robustness (DOR) values of the methods to the noise of missing data, genotyping error and phenocopy.

From: Performance analysis of novel methods for detecting epistasis

Noise Types

Models

Power

TEAM

BOOST

SNPRuler

AntEpiSeeker

epiMODE

Missing Data

Model 1

power 1

--

--

0.0000

1.0000

0.0000

  

power 2

--

--

0.0000

1.0000

0.0000

  

power 3

--

--

0.0000

1.0000

0.0000

 

Model 2

power 1

--

--

0.0000

0.0000

0.0000

  

power 2

--

--

0.0000

0.0000

0.0000

  

power 3

--

--

0.0000

0.0433

0.0000

 

Model 3

power 1

--

--

0.0000

1.0000

0.0000

  

power 2

--

--

0.2925

1.0000

0.0000

  

power 3

--

--

0.5188

1.0000

0.0000

 

Model 4

power 1

--

--

1.0000

1.0000

0.0000

  

power 2

--

--

1.0000

1.0000

0.0000

  

power 3

--

--

1.0000

0.7607

0.0000

 

Model 5

power 1

--

--

1.0000

1.1601

0.0000

  

power 2

--

--

1.0000

1.1601

0.0000

  

power 3

--

--

1.0000

1.1601

0.0000

 

Model 6

power 1

--

--

1.0000

0.0164

0.0000

  

power 2

--

--

1.0000

0.0214

0.0000

  

power 3

--

--

1.0000

0.0214

0.0000

 

Model 7

power 1

--

--

1.0000

1.0000

0.0000

  

power 2

--

--

1.0000

1.0000

0.0000

  

power 3

--

--

1.0000

1.0000

0.0000

 

Model 8

power 1

--

--

0.8399

1.0000

0.0000

  

power 2

--

--

0.8399

1.0000

0.0000

  

power 3

--

--

0.8407

1.0000

0.0000

 

Model 9

power 1

--

--

1.0000

1.0000

0.0000

  

power 2

--

--

1.0000

1.0000

0.0000

  

power 3

--

--

1.0000

1.0000

0.0000

Genotyping Error

Model 1

power 1

1.1759

0.0000

0.0000

1.0000

0.0000

  

power 2

0.9153

0.0000

0.0000

1.0000

0.0000

  

power 3

0.9170

0.0000

0.0000

1.0000

0.0000

 

Model 2

power 1

0.1531

0.0000

0.0000

0.0000

0.0000

  

power 2

0.0000

0.0000

2.0000

0.0000

0.0000

  

power 3

0.0000

1.0000

2.0000

0.0014

0.0000

 

Model 3

power 1

1.0000

0.0007

0.0000

1.0000

0.0000

  

power 2

1.0000

0.0007

0.0000

1.0000

0.0000

  

power 3

1.0000

1.1768

0.0000

1.0000

0.0000

 

Model 4

power 1

0.0031

1.0000

1.0000

1.1614

0.0000

  

power 2

0.0031

1.0000

1.0000

1.1614

0.0000

  

power 3

0.0031

1.0000

1.0000

0.9191

0.0000

 

Model 5

power 1

0.0051

1.0000

1.0000

1.0000

0.0000

  

power 2

0.0051

1.0000

1.0000

1.0000

0.0000

  

power 3

0.0051

1.0000

1.0000

1.0000

0.0000

 

Model 6

power 1

0.8026

1.0000

1.0000

0.8415

0.0000

  

power 2

0.8026

1.0000

1.0000

0.8415

0.0000

  

power 3

0.8026

1.0000

1.0000

0.8415

0.0000

 

Model 7

power 1

0.0058

1.0000

0.8399

1.0000

0.0000

  

power 2

0.0058

1.0000

0.8399

1.0000

0.0000

  

power 3

0.0058

1.0000

0.8399

1.1650

0.0000

 

Model 8

power 1

0.0000

1.0000

1.0000

1.0000

0.0000

  

power 2

0.0000

1.0000

1.0000

1.0000

0.0000

  

power 3

0.0000

1.0000

1.0000

1.0000

0.0000

 

Model 9

power 1

0.5050

1.0000

0.6862

1.0000

0.0000

  

power 2

0.5050

1.0000

0.6862

1.0000

0.0000

  

power 3

0.5050

1.0000

0.6862

1.0000

0.0000

Phenocopy

Model 1

power 1

0.0000

0.0000

2.0000

0.0005

0.0000

  

power 2

0.0000

0.0000

2.0000

0.0027

0.0000

  

power 3

0.0001

0.0000

2.0000

0.0643

0.0000

 

Model 2

power 1

0.0000

0.0000

0.0000

0.0000

0.0000

  

power 2

0.0000

0.1531

2.0000

0.0000

0.0000

  

power 3

0.0000

0.5785

2.0000

0.0003

0.0000

 

Model 3

power 1

0.0014

0.0000

0.0000

0.0023

0.0000

  

power 2

0.0153

0.0000

0.0001

0.0261

0.0000

  

power 3

0.1096

0.0000

0.0013

0.1542

0.0000

 

Model 4

power 1

0.7763

1.0000

1.0000

1.0000

0.0000

  

power 2

0.7763

1.0000

1.0000

1.0000

0.0000

  

power 3

0.7763

1.0000

1.0000

0.9393

0.0000

 

Model 5

power 1

0.5485

0.8808

0.1096

0.8006

0.0000

  

power 2

0.5485

0.8808

0.1096

0.9195

0.0000

  

power 3

0.6171

1.0000

0.1153

0.9195

0.0000

 

Model 6

power 1

0.8513

1.0000

1.0000

1.0000

0.0000

  

power 2

0.8513

1.0000

1.0000

1.0000

0.0000

  

power 3

0.8513

1.0000

1.0000

1.0000

0.0000

 

Model 7

power 1

0.1207

0.9203

0.3906

0.8315

0.0000

  

power 2

0.1207

0.9203

0.3906

0.8315

0.0000

  

power 3

0.2275

0.9601

0.5445

1.0208

0.0000

 

Model 8

power 1

0.0000

1.0000

0.9195

1.0000

0.0000

  

power 2

0.0000

1.0000

0.9195

1.0000

0.0000

  

power 3

0.0124

1.0000

0.9000

1.0000

0.0000

 

Model 9

power 1

0.5050

1.0000

1.0403

1.0000

0.0000

  

power 2

0.5050

1.0000

1.0403

1.0000

0.0000

  

power 3

0.5050

1.0000

1.0403

1.0000

0.0000

  1. There are three types of noise added into datasets respectively. For each model with certain type of noise, each method has three DORs, since three forms of detection power are introduced. Theoretically, DORs are ranged from 0 to 1. However, realistically, there are some DORs (only a few) in table larger than 1. Most DORs with italic fonts are caused by detection power computation precision and their detection power differences are not greater than 0.01. DORs with bold fonts, which are only occurred on robustness analysis of SNPRuler to phenocopy on Model 1 and Model 2, are described and evaluated in detail.
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us