Learning Bayesian network models of the atherosclerosis network
Atherosclerosis is a lifelong disease characterized by the deposition of fatty substances in the walls of the arteries, which can eventually reduce blood flow causing cardiovascular disease, including heart attack and stroke. Cardiovascular disease is the main cause of death in developed countries. The progression of atherosclerosis is seldom determined by a single factor, but rather by a network of interacting genetic and environmental factors. Little is known about this network which we call the atherosclerosis network.
The aim of this project is to learn Bayesian network models of the atherosclerosis network to improve the diagnosis and treatment of the disease. For instance, such a model would help to hypothesize genetic markers for reliable diagnosis and genetic targets for developing new drugs. We aim to learn the Bayesian network models from patient data. Specifically, the data for each patient consist of thousands of measurements, including clinical, life-style, and gene expression data.
Bayesian networks arose in the 80s at the intersection of computer science and statistics. Since then, they have become one of the most popular paradigms for data analysis. Unfortunately, the existing algorithms for learning Bayesian networks from data are computationally unfeasible when each patient is characterized by more than a few hundreds measurements. Thus, this project aims to develop computationally feasible algorithms for dealing with the data described above.
Page responsible: Patrick Lambrix
Last updated: 2009-08-24