Grant: Travel Awards - Institute Community Support

Complex diseases, such as autoimmune disorders, mental health conditions, and cancers, pose significant challenges to individuals and healthcare systems. Predicting who is at risk for these diseases early is crucial for timely interventions and prevention strategies that can improve personal health and reduce disease burdens. However, traditional risk prediction models often fall short because they do not adapt well to diverse populations or account for the dynamic nature of health data that changes over time. These limitations hinder the ability to provide accurate, personalized, and inclusive risk predictions for different groups. This research aims to improve disease risk prediction by leveraging advanced artificial intelligence (AI) methods, including large language models (LLMs) and domain adaptation. By combining genomic data with clinical, lifestyle, and socioeconomic information, the study will create a more flexible and precise risk prediction framework. The project has two key objectives: (1) to develop a "disease foundation model" using genomic and phenotypic data from the UK Biobank, and (2) to combine domain adaptation techniques to transfer the knowledge gained to the Canadian Longitudinal Study on Aging (CLSA). This ensures the model can accurately predict disease risk across populations with diverse characteristics. By making risk prediction models more adaptable and inclusive, this research will enhance the ability to identify high-risk individuals early, leading to improved personal health outcomes and better disease prevention. At the population level, it will address health disparities by ensuring predictions are accurate across different demographics. For healthcare systems, this approach offers a pathway to more proactive and efficient care delivery, reducing costs and improving overall system effectiveness. This innovative work will support a future where disease prevention is more personalized, inclusive, and effective.