Major depressive disorder (MDD) is a common mental illness that affects more than 300 million people worldwide. Depression is the most prevalent mental illness among U.S. veterans and increases their risk of hospitalization and death. Also, the prevalence of depression in veterans is significantly higher than in general U.S. adult population. Available treatments are ineffective for many MDD patients, and there is an urgent need to obtain a better understanding of biological bases of MDD in order to develop novel therapeutic strategies. Genetic contribution to MDD is ~40%, and recent large-scale GWAS discovered 102 significant loci that are linked to MDD risk. However, as in other complex disorders, the majority of the MDD risk loci reside within non-coding regions of the genome and are predicted to alter the activity of gene regulatory elements (GREs; e.g., enhancers or promoters). In addition, the effect of a GRE on gene expression is frequently dependent on the cell and tissue type, making it a challenge to understand the functional impact of GWAS risk loci. Linking the disease risk variants with expression quantitative trait loci (eQTLs) and/or with epigenetic maps of GREs in relevant tissues/cell types yielded formidable results in other psychiatric diseases and holds promise for MDD. In addition, a systematic experimental validation of eQTLs/enhancers’ activity via massively parallel reporter assays (MPRAs) provides a powerful platform to decipher the functional impact of the disease risk variants. To date, many neurobiological studies of MDD have been concentrated on neuronal dysfunction. However, multiple studies implicate microglia (MG)—the immune cells of the brain—to the pathogenesis of this disorder. Our preliminary data shows that MDD genetic risk is reflected in gene expression changes in the blood- derived myeloid cells. Because of the substantial overlap in transcriptomes of these cells and MG, these data suggest that MG might play an important role in susceptibility to MDD. However, large-scale gene expression data in purified MG are not available to test this hypothesis. Our overarching goal is to discern MG dysfunction in MDD as well as to link this dysfunction to genetic predisposition to MDD by integrating MDD GWAS, MG eQTLs, and epigenetic maps of enhancers that are active in MG. In Aim1, we will use our novel fluorescence-activated nuclei sorting (FANS) protocol (which allows the separation of MG from autopsied human brain) to generate high-quality gene expression data in MG purified from a large cohort of MDD cases and controls (N~300). Importantly, the cohort includes specimens from the U.S. veteran population. We will also perform a case-control comparison of gene expression in MG, including testing if alterations in MDD are different in veterans vs. non-veterans, In Aim2, we will use gene expression data from Aim 1 to map MG eQTLs, and will integrate these eQTLs with MDD GWAS findings using rigorous computational app...