2010 Poster Sessions : Independent Component Analysis: Identifying Data-Driven Human Gene Modules

Student Name : Jesse Engreitz
Advisor : Russ Altman
Research Areas: Artificial Intelligence
Cellular physiology, including disease states and drug responses, results from the combined influences of many genes. Experimentalists have now sampled many conditions and cell types, contributing vast amounts of gene expression data that represent many biological processes. The expansion of public microarray databases such as the Gene Expression Omnibus (GEO) allows the use of intelligent data mining approaches to extract information about these biological processes in a data-driven manner. Using 9,395 human microarrays measuring over 20,000 genes, we use independent component analysis to identify functional gene modules, or sets of genes, that describe a wide range of biological functions. Modeling gene expression as a linear combination of these modules, we identify novel effects of the pre-clinical anticancer drug parthenolide.

Jesse is a coterminal masters student at Stanford University working towards a B.S. in Biomedical Computation and an M.S. in Bioengineering. His research interests include translational bioinformatics, content-based microarray search, and dimension reduction methods for gene expression data. He currently works under the joint direction of Professors Russ Altman and Michael Clarke.