We conducted leukocyte-specific analysis using the Cardiogenics dataset27 composed of 738 monocytes and 593 macrophages samples. For each probeset – imputed SNP pair, a simple linear regression was used to analyze the data separately for monocytes and macrophages: yi=α+βxi+εi,1≤i≤n,εi~N(0,σ2)where i is the subject index, x is the effective allele copy number, and yi is the covariates-adjusted, inverse-normal transformed gene expression. Significance of cis (SNP within ±1Mb of the closest transcript end) eQTL effects were quantified with a Wald test on the ordinary Least Squares (OLS) estimator of the coefficient β, obtained with R. The distribution of the Wald test P values under the null hypothesis of no correlation between genotype and gene expression was estimated by rerunning the same analysis on a null dataset obtained by permuting the expression samples identifiers. For additional monocyte eQTL analysis, we queried statistics from Fairfax et al.28 to validate findings in the Cardiogenics dataset.
