EXCEED THE SPACE PROVIDED. This proposal details a plan for the candidate, currently a biomedical imaging specialist, to receive training and research experience in molecular biology and genomics in order to become an independent investigator in this field. The program begins with a two-year period of education and research training experiences designed to give the candidate exposure to the field of genomics in general, and specifically to microarray technology. The research part of the program will continue through the full five-year term of the grant, with half-time devoted during the first two years and nearly full time during the last three. The research program focuses on the analysis of data from microarrays, as applied to diagnosis and measurement in human disease applications. In the future, microarrays will be used diagnostically: tissue samples will be characterized by their gene expression patterns, and these patterns will be compared to the established expression patterns of a potentially large number of disease states. This presents challenges in efficient pattern analysis and in accurate and meaningful reporting of results. The research plan addresses these challenges in several ways. First, the candidate will develop software capable of generating simulated microarray images from an established set of a priori gene expression parameters. This project will continue through the entire term of the grant, beginning simply and evolving and becoming more complex and more accurate as the candidate progresses through the education plan. In parallel, the candidate will develop statistical models for the components of a specific cDNA microarray system, and these models will be incorporated into the simulation. In the third aim of the project, the candidate will develop efficient hierarchical descriptions and pattern analysis techniques for the problem of matching a sample expression pattern with one or several in a large set of possible classes. In the fourth aim, these techniques will be evaluated against an objective standard obtained from simulations, where the true classes are known. Finally, the simulation and classification methods will be applied to real problems in environmental toxicology and breast cancer. PERFORMANCE SITE ========================================Section End===========================================