Skip to main content

Table 2 Realized βa, mean bias, mean standard error (SE), coverage probability and percentage of the effects that had a p-value < 0.05 in the simulation where haplotypes with a rare allele at locus 2 had a modeled parameter estimate of 0.69. Each replicate contained 200 individuals

From: Estimating effects of rare haplotypes on failure time using a penalized Cox proportional hazards regression model

No penalty

Haplotype

Frequency

Realizeda

Mean bias

Mean SE

Coverage

P < 0.05b

010

0.020

0.741

0.014

0.44

0.95

0.43

011

0.005

0.593

-0.663

2971.81

0.95

0.21

110

0.002

0.000

-0.971

14514.13

0.96

0.10

111

0.280

0.713

0.018

0.09

0.79

1.00

001

0.050

0.012

0.050

0.28

0.91

0.09

100

0.020

0.038

0.010

33.23

0.95

0.05

101

0.003

0.000

-2.212

7674.81

0.90

0.10

Ridge penalty

Haplotype

Frequency

Realizeda

Mean bias

Mean SE

Coverage

P < 0.05b

010

0.020

0.741

-0.135

0.73

0.85

0.18

011

0.005

0.593

-0.359

1.91

0.74

0.03

110

0.002

0.000

-0.557

1.00

0.39

0.19

111

0.280

0.713

-0.010

0.11

0.83

1.00

001

0.050

0.012

0.037

0.24

0.87

0.13

100

0.020

0.038

0.070

1.08

1.00

0.00

101

0.003

0.000

-0.649

2.16

0.89

0.11

Difference penalty

Haplotype

Frequency

Realizeda

Mean bias

Mean SE

Coverage

P < 0.05b

010

0.020

0.741

-0.008

0.62

0.98

0.35

011

0.005

0.593

-0.132

1.39

1.00

0.03

110

0.002

0.000

-0.272

1.00

1.00

0.15

111

0.280

0.713

0.008

0.12

0.89

1.00

001

0.050

0.012

0.122

0.24

0.79

0.21

100

0.020

0.038

0.131

0.88

0.93

0.07

101

0.003

0.000

-0.301

1.68

0.91

0.09

  1. a) "Realized" denotes to the median estimate from the simulated data, where the haplotypes were known, and could be used in the Cox model directly; b) For haplotypes 010, 011, 110 and 111 these percentages reflect power, whereas the percentages for the other three haplotypes reflect type I error probabilities.