The data are simulated under the following linear mixed model :
$$ Y = X\beta + \omega + \varepsilon $$
with $omega ~ N(0, tau K)$ and
$epsilon ~ N(0, sigma^2 I_n)$.
The simulation uses $K$ only through its eigen decomposition; the parameter
K is therefore optional.
# generate a random positive matrix set.seed(1)
R <- random.pm(503)
# simulate data with a "polygenic component" y <- lmm.simu(0.3, 1, eigenK = R$eigen)
str(y)