Skip to main content

Table 3 Power to detect Gene × Environment (G×E) interactions.

From: New directions in childhood obesity research: how a comprehensive biorepository will allow better prediction of outcomes

  SAMPLE SIZE = 1000 SAMPLE SIZE = 2000
  G×E interaction OR = 2 G × E interaction OR = 2
Allele frequency (%) Binary environmental exposure prevalence (%) Binary environmental exposure prevalence (%)
  20 30 40 20 30 40
10 37.6% 45.8% 49.5% 64.2% 74.6% 78.6%
20 49.8% 58.9% 62.2% 79.0% 87.1% 89.4%
30 50.4% 58.7% 61.2% 79.6% 87.0% 88.7%
  G × E interaction OR = 2.5 G × E interaction OR = 2.5
Allele frequency (%) Binary environmental exposure prevalence Binary environmental exposure prevalence
  20 30 40 20 30 40
10 57.2% 67.9% 72.3% 85.7% 92.9% 95.0%
20 72.6% 81.7% 84.3% 95.2% 98.2% 98.7%
30 73.3% 81.3% 83.1% 95.5% 98.1% 98.5%
  G × E interaction OR = 3 G × E interaction OR = 3
Allele frequency (%) Binary environmental exposure prevalence Binary environmental exposure prevalence
  20 30 40 20 30 40
10 71.4% 81.9% 85.7% 94.7% 98.2% 99.0%
20 85.9% 92.5% 94.1% 99.0% 99.8% 99.9%
30 86.6% 92.2% 93.1% 99.1% 99.8% 99.8%
  1. OR = odds ratio. Disease prevalence is 25% throughout. Magnitude of associations between binary outcome and each of the gene and environmental risk factors are 1.6 throughout. Condition is assumed to be dominant throughout. Figures in bold represent scenarios where sufficient power for the study will be achieved.
Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us