Considering paired individuals among two datasets (iPSCs and HLCs), we further validated the tissue specificity of the iPSC and HLC cis-eQTLs by considering correlated data structure in the mixed effects model as implemented in Meta-Tissue (Sul et al., 2013). Three genotype PCs estimated from PCA of genotypes were used as covariates in mixed effect estimates calculations. As suggested by the developers, we subsequently used METASOFT (Han and Eskin, 2012) to estimate posterior probabilities (m-values) of tissue specificity incorporating correlations estimated by Meta-Tissue. In order to distinguish between iPSC and HLC eGenes identified via single tissue eQTL analyses, we used the m-value threshold of 0.9, as recommended by the developers, to determine the significance of the predicted cell-type effect. An eQTL-eGene pair was considered strongly associated with both cell types if m-value ≥ 0.9 in both cell types. Accordingly, an eQTL-eGene pair was considered HLC-specific if m-value ≥ 0.9 for HLCs and m-value < 0.1 for iPSCs. An eQTL-eGene pair was considered iPSC-specific if it had m-value ≥ 0.9 for iPSCs and m-value < 0.1 for HLCs. All other eQTL-eGene pairs
