Genetic risk scores of psychiatric phenotypes are associated with depression risk in a prospective Dutch...

Genetic risk scores hold potential for predicting depression in the general population. These scores must be validated for their associations with relevant characteristics of depression-related phenotypes, such as severity. We validated a genome-wide risk score (GRS) and a restricted polygenic risk score (PRS) for depression based on a meta-analysis of three genome-wide association studies and assessed their associations with depression in three subcohorts of middle-aged and older adults from the Dutch population-based Rotterdam Study.

Of participants with genotype data, 9,198 had longitudinally measured data (mean follow-up: 11.3 years) on three depression-related phenotypes (depressive symptoms, depressive syndrome, and major depressive disorder). Generalized linear models estimated the associations of standardized GRS and PRS with depression phenotypes per subcohort and were then meta-analyzed. One unit of the GRS/PRS represents 1 standard deviation, following z-transformation per cohort.

A one unit higher GRS and PRS were associated with any longitudinally measured depression phenotype (odds ratio (OR)_GRS = 1.20 [1.15-1.26], OR_PRS = 1.10 [1.05-1.16]). Effect sizes were highest for episodes of major depressive disorder: for individuals with the 10% highest GRS and PRS, the ORs were 1.99 [1.53-2.57] and 1.51 [1.13-1.99], respectively, compared to the middle 50% of the distribution.

The GRS and PRS for depression showed modest associations across multiple depression-related phenotypes in a population-based setting. The strength of associations generally increased with the severity of the phenotype. While effect sizes were generally larger for GRS compared to PRS, the difference was mostly not statistically significant.