ABSTRACT
This chapter provides a predictive-based perspective on the scientific issues surrounding the generation and successful use of toxicogenomics in hopes of continuing and refining the debate surrounding the use of gene expression data. It reviews analysis techniques integral to a predictive program, as are the parameters that affect the construction and utility of predictive toxicogenomics models. The chapter discusses the ToxExpress™ program, Gene Logic’s proprietary approach to toxicogenomics and presents specific applications of toxicogenomics using the ToxExpress reference database, analysis tools, and visualizations. Microarray platforms provide for tens of thousands of gene expression measurements, where each of the genes being measured has an associated noise produced from the physical aspects of chip processing and from biological differences associated with individual specimens. By limiting the focus in predictive toxicology to only those genes that truly change due to toxicity, investigators can effectively increase the signal-to-noise ratio and improve the chance for successful classification of new compounds.
