Covariates included age, sex, birth year, generation (parent versus offspring), and the two-way interaction between generation and age, birth year and generation, and generation and sex. In addition, in order to account for any residual population structure within the Caucasian sample, the first 10 principal components from an Eigenstrat analysis (Price et al., 2006) of the SNP data were used as covariates. When the clinical phenotypes were regressed on the covariates alone, between 13% and 28% of the variance was accounted for (Table 1). All demographic factors (including two-way interactions) were statistically significant in these regressions. In only a few cases, however, did a principal component from the Eigenstrat analysis add significantly to the prediction, suggesting there was little residual population structure in this Caucasian sample.
