-----------------------------------------------------------

Statistics Seminar Series

Thursday, April 20, 2006 @ 11:30AM
Cullimore Hall, Room 611
New Jersey Institute of Technology

Exploration and Analysis of DNA Microarray Data

Dhammika Amaratunga, Ph.D.

Senior Research Fellow, Nonclinical Biostatistics

Johnson & Johnson Pharmaceutical Research and Development LLC

Raritan, NJ

(Joint work with Javier Cabrera, Rutgers University)

Abstract

The emergence of DNA microarray technology has heralded in a new era of functional genomics research, one in which scientists can monitor gene expression profiles, tens of thousands of genes at a time. Microarray experiments generate many megabytes of data and their analysis is often incredibly challenging due to many issues: very little replication, high within array variability, nonlinear between array variability, high experimental variability, sporadic outliers, spatial outliers, nonnormality, within gene heteroscedasticity, between gene heteroscedasticity, correlations among genes, multiplicity, etc. This has created an opportunity for some very interesting statistical work. In this seminar, we will present an example of a DNA microarray dataset, through which we will illustrate many of the statistical issues involved and the solutions that have so far been proposed, including some of our own recent work.