We meta-analyzed three eQTL datasets with data from the dorsolateral prefrontal cortex (DLFPC): PsychEncode (N = 1387)49, ROSMAP (N = 461)50, and COGA-INIA (N = 138). ROSMAP eQTL analysis50 was limited to participants of European descent only and obtained from AMP-AD Knowledge Portal (https://www.synapse.org/#!Synapse:syn3219045). A full set of PsychEncode cis-eQTL summary statistics without P-value threshold restriction was downloaded from the PsychENCODE Integrative Analysis website (http://resource.psychencode.org/)49. The PsychENCODE Integrative Analysis contained adult brain prefrontal cortex data of 1387 individuals from the PsychENCODE and the Genotype-Tissue Expression (GTEx, https://www.gtexportal.org) data. Briefly, PsychEncode consortium included data from gene expression matrix normalized using quantile normalization, followed by inverse quantile normalization to map to a standard normal distribution (and to remove outliers). In all QTL analyses, known confounding factors such as age, sex, population substructure, and technical covariates were corrected for, and unidentified confounding factors were minimized through principal component analysis or similar methods. Detailed information about the data collection and analysis process can be found in the original study41. Human post-mortem brain samples (COGA-INIA) samples were obtained from the New South Wales Tissue Resource
