Depression is frequently comorbid with other psychiatric and somatic disorders resulting in
extensive use of diverse pharmacological treatments, decreased quality of life and increased
mortality rate. However, comorbidities may inform us about the different biological pathways
leading to and active in depression and thus can guide personalized medicine. In this project we
aim to apply advanced machine learning methods with existing data sets from public health and
biobanks to stratify depression-related multimorbidities using complete clinical patient
trajectories, medication history, metabolic, environmental risk profile, social inequalities and GWAS data. These strata will allow the development of a multimorbidity-adjusted disability score (MADS) for quantitative comparison, for health risk assessment and the application of chemoinformatics methods to explore new and more efficient drug candidates that minimize side-effects, drug-drug interactions, and polypharmacy.