Core--Statistics

点击翻译按钮获取中文摘要

Component 5 of the Center is concerned with the development and application of statistical models and computational methods for analysis of large and complex genomic data sets, including gene global objective is to develop statistical methods that relate such data to physiological or clinical states and outcomes. This basic methodological research program will use the data sets arising from the cardiovascular program is a development setting and test-bed, as well as others from cancer and other studies. These developments are integrated into the genome technology, genetic analysis and information components of the Center, link into data analysis from the candidate gene analysis and SNP discovery components, and interface with the education component. The project will result in broadly applicable bioinformatics tools for gene expression profiling, molecular phenotyping and screening, for large-scale association studies in relating large-scale SNP genotyping information to clinical outcomes and conditions, and for gene interaction analysis. Additional aspects involve developments in data basing and public domain implementations of novel statistical methodologies for genomic data. At the heart of the expression studies lies the guiding concept that it is the elucidation of structure and changes in expression patterns among possibly many interacting genes that together relate to outcomes in complex human diseases, even through many genes within the subset may exhibit small or modest expression levels and changes when considered in isolation. It is thus the aggregate effect of collections of genes and their interactions that are of fundamental concern. A major research priority is thus the development of methods to relate the high-dimensional expression profiles to defined outcomes and thereby deduce relevant patterns and implicated subset of genes. A complementary priority is then the development of efficient statistical and computational methods for the analysis of association of high-throughput SNP genotype data for such subsets of genes, and for others identified via other Center methods, again guided by the view that determining factors will often involve the combination and interaction of multiple gene variants.