Clinical stratification of Major Depressive Disorder in the UK Biobank: A gene-environment-brain Topological Data Analysis

Major depressive disorder (MDD) is a leading cause of disability worldwide, affecting over 300 million people and posing a significant burden on healthcare systems. MDD is highly heterogeneous, with variations in symptoms, treatment response, and comorbidities that could be determined by diverse etiologic mechanisms, including genetic and neural substrates, and societal factors. Characterizing MDD subtypes with distinct clinical manifestations could improve patient care through targeted personalized interventions. Recently, Topological Data Analysis (TDA) has emerged as a promising tool for identifying homogeneous subgroups of diverse medical conditions and key disease markers, reducing complex data into comprehensible representations and capturing essential dataset features. Our study applied TDA to data from the UK Biobank MDD subcohort composed of 3052 samples, leveraging genetic, environmental, and neuroimaging data to stratify MDD into clinically meaningful subtypes. TDA graphs were built from unimodal and multimodal feature sets and quantitatively compared based on their capability to predict depression severity, physical comorbidities, and treatment response outcomes. Our findings showed a key role of the environment in determining the severity of depressive symptoms. Comorbid medical conditions of MDD were best predicted by brain imaging characteristics, while brain functional patterns resulted the best predictors of treatment response profiles. Our results suggest that considering genetic, environmental, and brain characteristics is essential to characterize the heterogeneity of MDD, providing avenues for the definition of robust markers of health outcomes in MDD.