EcoGenomics: analysis of complex systems via fractal geometry

Ecogenomics is a convenient descriptor for the application of advanced molecular technologies to studies of organismal responses to environmental challenges in their natural settings. The development of molecular tools to survey changes in the transcript profile of thousands of genes has presented scientists with enormous analytical challenges. In the main, these center about the reduction of massively paralleled data to statistics or indices comprehensible to the human mind. Historically, scientists have used linear statistics such as ANOVA to accomplish this task, but the sheer volume of information available from microarrays severely limits this approach. In addition, important information in microarrays may not reside solely in the up or down regulation of individual genes, but rather in their dynamic, and probably nonlinear, interactions. In this presentation, we will explore alternative approaches to extracting of these signals using artificial neural networks and fractal geometry. The goal is to produce predictive models of gene dynamics in individuals and populations under environmental stress and reduce the number of genes that must be surveyed in order to recover transcript profile patterns of environmental challenges.