|Grant Number:||5R01CA106320-08 Interpret this number|
|Primary Investigator:||Zhao, Lue Ping|
|Organization:||Fred Hutchinson Cancer Research Center|
|Project Title:||Methods for Correlating Snp Haplotypes with Cancer|
DESCRIPTION (provided by applicant): The long-term objectives of our research efforts are to develop innovative research design methods and analytic methods for the study of human diseases in the context of maturing genomic technologies (genetic polymorphisms, gene expressions and protein expressions). The primary aims of this proposal are to develop statistical methods and study designs to objectively assess the relationship of genetic polymorphisms in the form of SNPs and haplotypes with disease phenotypes and to investigate the role of gene-environment interactions, believed to play a major role in disease etiology. This objective represents the continuation of our current research project, which focuses on mapping complex diseases. Recent developments in genotyping technologies are revealing interesting local haplotype structures and are continuing to discover many more genetic markers throughout the genome. These provide not only unprecedented opportunities to gain a deeper understanding of disease etiology and to discover biomarkers useful for disease prevention and control, but also present a considerable challenge to the extraction of useful information from the ocean of genomic data now available. While extended pedigrees, nuclear families, sib-pairs have been used in indirect association analyses, direct association analyses may be more efficient and yield sufficient power to detect gene-environment interactions. Therefore, to maximize the power of the association analysis, case-control designs (both matched and unmatched) and cohort study designs will be applied. Methods to be developed will be based on well-established statistical techniques, including: genetic models, estimating equation techniques, likelihood methods, and logistic regression techniques. Where analytical solutions are not possible or computationally prohibitive, bootstrap techniques and Monte Carlo simulation methods will be applied to estimate relevant statistics. A range of statistical methods will be developed, including novel algorithms as necessary, and implemented for dissemination to the research community. Then these methods will be applied data from actual studies for validation and illustration purposes. To ensure the integration of statistics, biology and epidemiology, a team of investigators from multi-disciplinary backgrounds will discuss issues and jointly generate and critique statistical solutions.