Statistical Modeling of Graph Theoretic Data in Systems Biology

Scholtens 

Wednesdays@NICO Seminar, Noon, April 08 2009, Chambers Hall, Lower Level

Prof. Denise Scholtens, Northwestern University

Abstract 

Node-and-edge graphs are a foundational structure for recording, visualizing and analyzing high-throughput genomics and proteomics data. Like most data, systems biology observations generated by high-throughput technologies are subject to measurement error and therefore must be treated accordingly. Frequently reported summary statistics for these data often fail to account for experimental design and the stochastic nature of the observations. We apply classic statistical modeling approaches for a variety of problems, thereby improving inference on commonly reported graph statistics, local features of interest in global graphs, and plausible error probability bounds.