Cardiovascular disease (CVD) is a leading global health burden and often accompanied by complications associated with the kidney, liver, and brain. The complex interplay between CVD and extra-cardiac organs accelerates disease progression and impairs clinical management, yet the underlying bidirectional mechanisms remain poorly understood. This project aims to understand systemic organ crosstalk in CVD using the UK Biobank’s extensive multimodal dataset.
We will employ a combination of machine learning and causal inference approaches to identify multi-organ interactions. Specifically, we will integrate clinical data, proteomics, and metabolomics to build predictive models for organ-specific complications. Mendelian randomization will be used to assess causal relationships between genetic variants and organ dysfunction outcomes. Models will be validated across age-, sex-, and race-stratified cohorts to ensure robustness and generalizability.
This three-year study is expected to provide mechanistic insights into systemic drivers of CVD-related multiorgan dysfunction and identify high-risk individuals who may benefit from organ-protective interventions. The findings may inform integrated clinical strategies and improve outcomes for patients with complex chronic disease.
We will provide updates on the progress and outcomes of our research project using UK Biobank in annual report. And we plan to disseminate our research findings through publication in peer-reviewed scientific journals.
In accordance with UK Biobank’s AI policy, we will follow prevailing standards for responsible AI when we undertake research using AI. All trained parameters derived from UK Biobank data will be returned as part of our results data submission. We will not use UK Biobank participant-level data in any public generative AI models, nor upload any part of the dataset or derived outputs to public repositories such as GitHub.
