|UM SPH Home > Faculty & Research > Epidemiology > Baylin Group - Nutritional Epidemiology and Cardiovascular Disease > Research > Toward a New Molecular Classification of Cancer|
Baylin Group - Nutritional Epidemiology and Cardiovascular Disease
Toward a New Molecular Classification of Cancer
National Cancer Institute (NCI)
About this Project
For most types of cancer histopathology is insufficient to predict disease progression and clinical outcome. Working with researchers at the UM Cancer Center, we helped develop and apply methods for identifying genes expression profiles based on microarray analysis that could be used for identifying histological subtypes of cancers, distinguishing between cancer stages, and predicting patient survival. We have used a wide range of methods to study those gene expression patterns including machine-learning algorithms, scan statistics, markov-chain monte carlo methods, and risk index methods.
Image from Beer, et al. Nature Medicine 8, 816 - 824 (2002)