This week's Mathematical Sciences Department colloquium is given by Dr. Min Chen, Assistant Professor of Co Division of Biostatistics, Department of Clinical Sciences, University of Texas Southwestern Medical Center.
In genome-wide association studies (GWAS) researchers examine a large number of markers to identify their associations with disease, or to prioritize markers for follow-up studies. In most published studies the search is limited to single markers. However, this approach may lack adequate statistical power for true discoveries. We propose to incorporate biological pathway information in GWAS by a Markov Random Field model. This is motivated by the observation that genes interact with each other and multiple genetic markers may jointly affect the disease risk. Besides, a large amount of knowledge about biological pathways and gene-gene interactions has been accumulated from past biological and bioinformatics studies. Unlike most existing methods that treat genes in a pathway as an exchangeable gene list, our approach takes into account functional relationships among those genes. We show that the conditional distribution of our MRF model takes on a simple form, and propose an iterated conditional modes algorithm as well as a decision theoretic approach for statistical inference of each gene’s association with disease.
Coffee will be served in FO 2.610F at 1:30 PM.
Please visit our webpage at http://www.utdallas.edu/math/news/ for up-to-date colloquia information.