Statistical Techniques for Microarray Data: A Partial Overview

With the advent of the modern microarray technology, biologists can now observe the simultaneous (joint) expression profiles of thousands of genes in a single experiment. The large volume of data generated from these experiment (many available on public Internet sites) has created tremendous opportunities for the statisticians to get involved in this exciting development and create appropriate statistical tools for analyzing these data. As pointed out recently by Terry Speed (2001 Joint Statistical Meeting, Wald Lecture III), the number of papers involving microarray data seems to be exponentially growing over the past few years. We have attempted to present an overview of some of the major statistical developments in this area. They include statistical clustering, normalization and correction for systematic bias, empirical Bayes testing of significance of differential gene expression, ANOVA models and exploration of relationships of expression levels between genes. The scope of such an overview is ‘partial’ by nature because new and novel statistical papers dealing with statistical analysis of microarray data are being added to the literature at an increasing pace. S. Datta Statistical Techniques for Microarray Data 2