Gene expression simulation settings.

Fourteen different settings under which we generated transciptomics data sets. Setting differed in sample size (N), number of clusters (g), number of fixed effects (p), number of random effects (q), correlation in auto-regressive covariance matrix (ρ), number of non-zero fixed effects (# effects) and random effects covariance (Ψθ) structure. For values of the parameters, see Table A in S1 Tables. Setting 2 is identical to setting “H2” in [13] and the parameters used are identical.