To quantitatively compare the SCZ GWAS and eQTL association signatures, we modified the R package coloc39 for Bayesian inference of co-localization between the two sets of summary statistics across each gene’s cis-region. Coloc2, our modified implementation of coloc, analyzes the hierarchical model of gwas-pw,43 with likelihood-based estimation of dataset-wide probabilities of five hypotheses (H0, no association; H1, GWAS association only; H2, eQTL association only; H3, both but not co-localized; and H4, both and co-localized). We then used these probabilities as priors to calculate empirical Bayesian posterior probabilities for the five hypotheses for each locus, in particular PPH4 for co-localization.