Dynamic risk prediction of complex diseases
Approved Research ID: 71274
Approval date: February 7th 2022
Complex diseases such as cancer often manifest later in life as the result of genetic predispositions, environmental factors, and their interactions. An individual's genetic factors can inform their baseline risks to develop different diseases, but these risks are modulated by environmental factors as the individual ages. Medical records contain information that may predict future disease onset, so this information can be used to update disease risks over time. However, genetic data and longitudinal medical data are very different types of data, and there are currently no methods to integrate them effectively. We therefore aim to develop statistical models and computational algorithms for this integrative analysis, using both deep learning and Bayesian hierarchical modeling approaches. Above all, this project will establish an integrative method for dynamic risk prediction that may be used to help patients understand how their disease risks change over time based on their multitude of genetic and medical factors.