Statistical Genetic Methods for Cancer Gene Studies

DESCRIPTION (provided by applicant): Finding the genetic determinants of common complex disorders is a formidable task, since many genetic and environmental factors, either individually or in combination, influence trait expression. The use of association studies to tackle this challenge is becoming more widespread, especially with the release of the human genome sequence. However, large-scale association studies with sufficient power will be required to find these genes with modest effect sizes. Family-based association studies have consistently been shown to be less powerful than case-control studies, but case-control studies do not efficiently use the data collected for linkage studies and national registries. Therefore, the overall goal of this mentored proposal is to develop and investigate novel study designs and statistical analysis methods for association studies that are a combination of the two aforementioned designs. Specifically, we will: (1) examine sampling strategies for our proposed hybrid design in order to provide guidelines that enable researchers to choose the most informative subjects. (2) derive and evaluate statistical methods to be used for our proposed hybrid design. We will consider methods for stratified analyses, for matched designs, for DNA pooling, and for testing interaction effects. (3) develop and examine statistical methods that correct for population stratification, since our proposed designs may be susceptible to the effects of population stratification. (4) develop user-friendly software that can be used for the design and analysis of our proposed hybrid designs. Additionally, through the planned didactic training in areas including statistical genetics and molecular biology, this mentored proposal will provide me the opportunity to build a strong foundation upon which to build career conducting methodological research involving genetic studies of cancer.