Polygenic Risk for Major Depression: Diagnostic Specificity and Developmental Trajectories in an Admixed Youth Cohort

Major depressive disorder (MDD) is heritable and polygenic, yet the relative diagnostic specificity, developmental impact, and cross-ancestry generalizability of MDD polygenic risk scores (MDD-PRSs) remain unclear.

At 2 sites of the Brazilian High-Risk Cohort (N = 2163; ages 6-23 years; 45.6% female), we used a discovery-replication design to test associations between MDD-PRSs and 1) lifetime DSM-IV diagnoses, 2) longitudinal depressive-symptom trajectories derived from latent growth-curve models, and 3) late-adolescent/young-adult depressive symptoms and nonsuicidal self-injury (NSSI). Analyses used weighted ordinary least squares and quantile regression, adjusting for age, sex, socioeconomic status, genetic principal components, and psychiatric comorbidity.

MDD-PRSs showed relative specificity for MDD in both sites, explaining approximately 2.6% to 3.6% of the liability-scale variance; no other diagnostic category survived false discovery rate correction. Longitudinally, higher MDD-PRSs predicted a higher initial level (β = 0.10-0.18; p < .001 in both sites), a faster rate of increase (β = 0.07-0.10; p = .004 and p < .001), and a higher time-averaged level (β = 0.13-0.20; p < .001 in both sites) of depressive symptoms, with effects strengthening at higher quantiles. Cross-sectionally, MDD-PRSs were associated with more depressive symptoms (β = 0.15-0.17; p < .001 in both sites). For NSSI, a main-effect association was observed (β = 0.09; p < .001), but it remained significant only in a post hoc MDD-PRS × lifetime-MDD interaction among individuals with MDD (β = 0.24; p = .014).

In this deeply phenotyped, admixed cohort, MDD-PRSs showed relative diagnostic specificity and marked a more severe developmental course of depression, with NSSI associations contingent on MDD diagnosis, supporting the relative specificity, developmental utility, and cross-ancestry relevance of MDD-PRSs.