Skip to main content
. 2014 Jul 14;16(1):143–154. doi: 10.1093/biostatistics/kxu034

Table 1.

Statistical significance for the test of calibration of alternative models for joint effects of SNPs for breast cancer in the BPC3 study

Multiplicative model
 Additive model
Complete case analysis
 Analysis including subjects with missing genotypes
 Complete case analysis
Inline graphic OR Inline graphic OR Inline graphic OR Inline graphic OR Inline graphic OR Inline graphic OR
HL test 0.11 0.87 — — 0.0003 0.01
Asymptotic Inline graphic 0.11 0.85 0.16 0.11 Inline graphic Inline graphic
Inline graphic 0.20 0.77 0.23 0.17 Inline graphic Inline graphic
Parametric bootstrap Inline graphic 0.07 0.70 0.14 0.12 Inline graphic Inline graphic
Inline graphic 0.11 0.54 0.28 0.18 Inline graphic Inline graphic

The proposed method is evaluated using Inline graphic or Inline graphic with the corresponding risk regions being defined by the combinations of evenly placed grid points at various upper and lower quintiles of the risk distribution. Inline graphic are based on Inline graphic simulations. Inline graphic are based on Inline graphic simulations. Inline graphic OR is the analysis assuming a homogeneous OR across cohorts. Inline graphic OR is the analysis assuming heterogeneous OR across cohorts.