THE IDENTICAL DEPRESSION PHENOTYPING CONSORTIUM: DECONSTRUCTION AND PREDICTION OF MDD AND TREATMENT RESPONSE

The Identical Depression Phenotyping Consortium consists of studies in the UK (Genetic Links to Anxiety and Depression or GLAD and UK Biobank), the Australian Genetics of Depression study, and the Biobanks Netherlands Internet Collaboration (BIONIC). The three studies are using the same method of phenotyping depression with detailed demographics, clinical record linkage, and data on over 130,000 cases of Major Depressive Disorder. We propose a symposium focused on advancing predictive models in MDD and its treatment, emphasizing the integration of polygenic scores, family history, and clinical data.

Wang will present on Joint Multi-Family History and Multi-Polygenic Score Prediction of Major Depressive Disorder. Machine learning integrating these factors in GLAD (9,927 MDD cases, 4,452 controls) revealed significant prediction accuracies for MDD, the number of recurrent MDD episodes. These findings were replicated in UK Biobank (40,667 MDD cases, 70,755 controls). Next, Li will present on incorporating genetic and clinical predictors for antidepressant side effects in > 5K cases from the GLAD study. By employing machine learning models, they achieved significant success in predicting side effects and discontinuation rates, particularly when integrating data from prior prescriptions. Huider will present on genetic analyses of MDD on behalf of the BIONIC consortium presents a large-scale genetic analyses of MDD and its symptoms to explore depression heterogeneity within the Netherlands, utilizing uniform in-depth phenotyping in > 30K cases. This ambitious project highlights the importance of large, homogeneous datasets in deciphering the complex genetics of depression. Finally, Mitchell will present on Using polygenic risk scores to characterise treatment resistant MDD in to explore the association of TRD with biological predictors such a polygenic score (PGS) and CYP2C19 and CYP2D16 metaboliser profiles, measured personality traits, and environmental predictors such as social support and exposure to stressful life events. Lastly, they tested for any gene-environment interactions across predictors. Their research identifies genetic factors that correlate with long-term treatment outcomes, providing a basis for personalized medicine in treating depression.

This symposium aims to showcase cutting-edge research that integrates genetic, familial, and clinical data to predict and manage major depressive disorder more effectively. Discussant Hatoum will consider the implications of integration of genetic prediction with machine learning approaches and the possibilities for clinical utility.