Briefly, y is the vector of log-transformed lymphocyte counts. X is a matrix of covariates (age and an intercept term in this model) and β is a vector of the fixed-effect coefficients. G reflects the number of copies of an allele at a given SNP, while γ, the main quantity of interest in this model, is the effect size of each additional copy of that allele (we consider an additive model at each SNP). g is a random effect accounting for the polygenic background, which is assumed to be multivariate normally distributed with mean 0 and covariance determined by the kinship coefficient, additive genetic variance and dominance genetic variance. e accounts for the residual (or environmental) error and is assumed to be normally distributed with mean 0 and variance. The two random effects are weighted by the respective variance component estimates from the first step of the analysis (determining heritability). The plots shown in Fig. 2B and C were originally created using LocusZoom (75).
