Hotel Estoril Eden, Monte Estoril,
5-8 October 2005



NextText Box: Participants
Text Box: Programme

esign of Genetical Genomics Studies Which Use Two-Color Microarrays

Steven G. Gilmour1, Júlio  S. S. Bueno Filho2 and Guilherme J. M. Rosa3
School of Mathematical Sciences, Queen Mary, University of London, UK
Universidade Federal de Lavras, Brazil
Michigan State University, USA

Microarray experiments have been used extensively to provide a surrogate measure for comparison of gene activity in different tissues, cell types, developmental stages, or experimental conditions. The application of this technology however, especially for two-colour platforms, poses some experimental design challenges. As samples are assayed in a pairwise competitive hybridization fashion and with different dye labelling on each slide, microarray experiments using two-color systems define row-column, or two-way blocking, designs.

Most papers on design for two-color microarray experiments focus on the comparison of a discrete set of designs, such as reference and circular structures, in situations where treatment effects are assumed fixed, and generally without any structure. Microarray experiments however have been used recently in genetical genomics studies, as an additional tool to understand the genetic mechanisms governing variation in complex traits, such as for estimating heritabilities of mRNA transcript abundances, for mapping expression quantitative trait loci, and for inferring regulatory networks controlling gene expression.

This new application of gene expression assays greatly broadens the concept of treatments (or experimental groups) in microarray experiments. Under these circumstances, treatments may now refer to families or to subjects (within family structures or complex pedigrees) in a given experiment. In these cases treatments are more appropriately considered to be random effects, with specific covariance structures. In this presentation, we discuss the optimal design of two-colour microarray experiments whose goals refer to specific genetic questions. Some examples are used to illustrate the choice of a design for comparing fixed, structured treatments, such as genotypic groups.