QuBBD: Large Scale Modeling of Big Multi-cohort Data for Cardiovascular Diseases and Type 2 Diabetes

Brown University, Providence RI

Investigators

Abstract

Cardiovascular disease and Type 2 diabetes are leading causes of death in the United States. It is well known that these diseases are the result of multiple genetic, environmental, and lifestyle risk factors. Understanding the complex interplay of these risk factors will lead to identification of key and causal risk factors that will provide the foundation for developing personalized prevention and treatment strategies. To this end, decades of research efforts have focused on the collection of comprehensive data from large populations. However, many challenges remain in how to model and uncover the complex disease mechanisms from such big datasets, and thus, gaps remain in how to best diagnose, treat, and prevent these diseases. This award supports initiation of a collaborative research project that brings together cardiovascular/diabetes scientists and statisticians to address these challenges. The project addresses the development of large-scale mathematical models of genetic, environmental, and lifestyle risk factors, using existing datasets from three reputable national studies, including the Framingham Heart Study, the Jackson Heart Study, and the Women's Health Initiative. This project will first develop large-scale network models for high dimensional risk factors. Specifically, this project will develop large-scale and computationally feasible graphical models for heterogeneous data with mixed data types and complex dependence, and then employ graph theoretical approaches to identify the key drivers. Based on the identified networks and key drivers, novel and interpretable predictive rules will be developed. These modeling results will be validated systematically using multi-cohort data. Finally, this project will deploy a web-based portal for disseminating the modeling results, which will serve as a core tool for research and learning. This award is supported by the National Institutes of Health Big Data to Knowledge (BD2K) Initiative in partnership with the National Science Foundation Division of Mathematical Sciences.