Objectives: The objectives of this study were to investigate the effect of genetic and social factors on depressive symptoms and depression over time and to test whether social factors moderate the relationship between depressive symptoms and its underlying genetics in later life. Methods: The study included 2,279 participants with a mean follow-up of 15 years from the Longitudinal Aging Study Amsterdam with genotyping data. The personal genetic loading for depression was estimated for each participant by calculating a polygenic risk scores (PRS-D), based on 23,032 single nucleotide polymorphisms associated with major depression in a large genome-wide association study. Partner status, network size, received and given emotional support were assessed via questionnaires and depressive symptoms were assessed using the CES-D Scale. A CES-D Scale of 16 and higher was considered as clinically relevant depression. Results: Higher PRS-D was associated with more depressive symptoms whereas having a partner and having a larger network size were independently associated with less depressive symptoms. After extra adjustment for education, cognitive function and functional limitations, giving more emotional support was also associated with less depressive symptoms. No evidence for gene-environment interaction between PRS-D and social factors was found. Similar results were found for clinically relevant depression. Conclusion: Genetic and social factors are independently associated with depressive symptoms over time in older adults. Strategies that boost social functioning should be encouraged in the general population of older adults regardless of the genetic liability for depression.

Progress in high-throughput metabolic profiling provides unprecedented opportunities to obtain insights into the effects of drugs on human metabolism. The Biobanking BioMolecular Research Infrastructure of the Netherlands has constructed an atlas of drug–metabolite associations for 87 commonly prescribed drugs and 150 clinically relevant plasma-based metabolites assessed by proton nuclear magnetic resonance. The atlas includes a meta-analysis of ten cohorts (18,873 persons) and uncovers 1,071 drug–metabolite associations after evaluation of confounders including co-treatment. We show that the effect estimates of statins on metabolites from the cross-sectional study are comparable to those from intervention and genetic observational studies. Further data integration links proton pump inhibitors to circulating metabolites, liver function, hepatic steatosis and the gut microbiome. Our atlas provides a tool for targeted experimental pharmaceutical research and clinical trials to improve drug efficacy, safety and repurposing. We provide a web-based resource for visualization of the atlas (http://bbmri.researchlumc.nl/atlas/).