Detection and Understanding of Expression Differentiatio

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Recent studies suggest that the regulation of longevity may be conserved in many eukaryotes ranging from yeast to mammals. In this project, we plan to obtain expression profiles on a time course for yeast strains with normal and extended life span, to develop novel statistical methods to detect expression differentiation, to develop new statistical and computational method to understand the pathways leading to extended life span, and to disseminate data and software resulting from the project. From microarray measurement, we seek differentiation of mRNA expressions among different biological samples. The statistical treatment of normalization aims to reduce army-specific "block effect" due to uncontrolled variation. To adjust for spatial patterns in both background and scale, we propose sub-array normalization. According to our experimental design, substantial differentiation may exist among arrays and we detect them by the technique of least trimmed squares, whose exact solution can be computed by a fast and stable algorithm we developed recently. From microarray analysis of strains such as sch9? and ras2? mutants, those genes with significant differentiation will serve as seeds for future investigation. We will pursue several directions. First, we search for genes that are co-expressed with seeds. Second, we search for changes of m-expressions associated with seeds. Third, we investigate regulation related to seed genes. Fourth, using the protein-protein interaction databases available for growing yeast we attempt to identify protein complexes as well as pathways that regulate longevity in yeast. Many genomic or interaction data such as protein-DNA interaction data can be arranged in a binary array. We introduce the structure of directed acyclic Boolean (DAB) networks as a tool of exploring biological pathways from binary arrays. With a few reasonable starting networks, we wilt use more sophisticated Bayesian networks to polish and refine the final results. This project aims to discover causations of life span extension from systematic experiments of yeast expression. To serve the scientific mission, we develop and integrate statistical and computational methodologies. Our research will benefit the society by new understanding of aging.