Decoding Depression: Integrating Genetics, Environment, and Clinical Data to Predict Prognosis and Treatments

Major depressive disorder (MDD) is a prevalent psychiatric condition with significant genetic components, often complicated by comorbidities that hinder an accurate diagnosis and effective pharmacological or psychological therapy. Typically, patients with MDD first seek help from general practitioners (GPs). While GPs can usually identify MDD, they may lack the specialized training required to detect additional conditions such as attention deficit hyperactivity disorder (ADHD), post-traumatic stress disorder (PTSD), eating disorders, or obsessive-compulsive disorder (OCD). These conditions often necessitate different treatment approaches. If these conditions are not correctly identified or are overlooked, it can result in less effective treatment and an increased risk of side effects. Given the genetic links between MDD and these frequently comorbid conditions, integrating their genetic information with social, environmental, and clinical data could significantly enhance personalized treatment strategies, tailored to individuals' unique genetic and clinical profiles. Such personalized care could improve treatment outcomes and mitigate the burden of comorbid conditions, which are often challenging to manage alongside MDD. To address these challenges, I propose a comprehensive three-phase research strategy that integrates genetic data with clinical and social information, applied to the largest cohort worldwide focused on the genetics of severe depression and anxiety: Mapping the Genetic and Phenotypic Landscape: Analyze self-reported questionnaires, mental health diagnoses, and medical records to elucidate how different conditions coexist with MDD and affect treatment responses. Employ Genome-Wide Structural Equation Modelling (G-SEM and GW-SEM) to explore complex genetic relationships between MDD and its comorbidities, identifying specific genetic markers associated with MDD, both with and without other conditions. Predictive Modelling: Building on the genetic data collected, I will develop polygenic risk scores (PRS) for depression and its common comorbid conditions. These scores will be combined with clinical data, family history, and socioeconomic information using advanced machine learning models. This integrated approach aims to enhance the prediction of treatment responses, side effects, and clinical trajectories, providing more precise and effective treatment options. Validation and Expansion: The predictive models will be rigorously tested using data from diverse external cohorts, such as PRADA and Genes & Health. This validation process will assess the generalisability and robustness of the models across different populations and ancestries. Throughout this project, I intend to disseminate our findings at major scientific conferences, collaborate with patient advocacy groups, and ensure open access to publications and the open-source code for the developed tools. I will also engage regularly with clinicians and mental health professionals who have expertise in mood disorders to ensure that the research findings have practical, real-world applications. Ultimately, this project aims to advance the field of personalised depression treatment by integrating genetic data with clinical insights. By examining MDD and its common comorbidities within diverse social contexts, I aim to develop a deeper understanding of how these conditions influence treatment responses, clinical trajectories, and therapy adherence. The goal is to create robust, unbiased tools that benefit individuals across all social contexts and ethnic backgrounds, helping to reduce health disparities in depression treatment and improve outcomes for those affected by this complex and challenging disorder.