Research Outline:
This project aims to unravel the complex interrelationships between mental health, organ-level imaging phenotypes, systemic molecular profiles, and the development of multimorbidity.
Key Research Questions:
What are the cross-sectional and longitudinal associations between mental health status, brain/body imaging-derived phenotypes, proteomic/metabolomic biomarkers, and specific clusters of multimorbidity?
Do these associations reflect causal relationships? Can genetic instruments for mental health, protein levels (pQTLs), and metabolites (mQTLs) causally influence the risk of multimorbidity?
Can a multimodal model integrating mental health, imaging, proteomics, metabolomics, blood characteristics and genetics outperform conventional models in predicting future multimorbidity?
Objectives:
To identify and define multimorbidity clusters using linked health records.
To assess associations using regression models, incorporating mental health scores, imaging phenotypes (e.g., brain structure, cardiac function), and Olink proteomics/NMR metabolomics data.
To perform Mendelian Randomization analyses to infer causality between these multimodal exposures and multimorbidity outcomes.
To develop and validate an integrated risk prediction model for multimorbidity.
Scientific Rationale:
Mental health, brain and body imaging features, and plasma proteomic/metabolomic levels offer complementary biological insights into disease risk. Their integrated analysis is crucial for understanding the shared pathways leading to co-occurring chronic diseases (multimorbidity). The UK Biobank provides a unique platform with all these data modalities available in a large population cohort. This project will leverage this resource to move beyond single-disease studies, uncovering multimodal biomarkers and causal mechanisms for complex multimorbidity, ultimately aiding early risk stratification and targeted prevention.
