COLLOQUIUM

  DEPARTMENT OF MATHEMATICS AND STATISTICS

and

SEMINAR ON QUANTITATIVE BIOLOGY

DEPARTMENT OF PHYSICS

OAKLAND UNIVERSITY

Shyamal D. Peddada

Biostatistics Branch

National Institute of Environmental Health Sciences

Research Triangle Park, NC

Constrained Inference in Gene Expression Analysis

Abstract

Toxicologists are often interested in understanding changes in gene expression over “ordered conditions” such as amount of exposure to a chemical (dose-response), duration of exposure to a chemical (time-course) or tumor stages etc.  For example, one may be interested in identifying the expression profile/pattern of a gene during various stages of cancer and cluster genes with similar pattern.  Genes belonging to a particular pattern of expression may potentially have a biological interpretation.  It may not always be feasible to develop parametric models to determine patterns but often it is possible to use mathematical inequalities to describe patterns.  Thus, the problem of identifying genes with similar  pattern can be restated as a constrained inference problem where the mean vector of expression belongs to a convex cone in p-dimensional Euclidean space, where p is the number of ordered conditions.  In some instances inequality constraints may arise naturally on a unit circle instead of the p-dimensional Euclidean space.  For instance cell-cycle experiments are routinely conducted to determine, among other things the phase angle associated with each cell-cycle gene.  Thus, in this case the parameter space is described by points on a unit circle.  Based on available literature and known biological functions of a subset of cell-cycle genes, one may expect an order among the phase angles around the unit circle, thus inducing an isotropic order (constraint) among a subset of cell-cycle genes.  In this talk we shall describe some statistical methods for analyzing gene expression data which exploit such underlying constraints on the parameter space.  Some examples will be provided to illustrate the methodologies.

Thursday, March 18, 2010

3:00 – 4:00 P.M.

372 Science and Engineering Building

(Refreshments at 2:30-3:00 PM in the kitchen area adjacent to 368 SEB)