|Grant Number:||5R01CA052862-14 Interpret this number|
|Primary Investigator:||Thomas, Duncan|
|Organization:||University Of Southern California|
|Project Title:||Survival Models for Genetic Epidemiology|
DESCRIPTION (provided by applicant): This project is primarily concerned with gene characterization for censored age-at-onset disease traits like cancer. During previous cycles, we have investigated a number of design issues for family-based studies, such as those arising in the context of our work on the NCI's Cooperative Family Registries for Breast and Colorectal Cancer Studies (CFRBCCS) as well as methods for modeling penetrance in relation to age and multiple genetic and environmental factors. The present proposal represents a continuation of work on these major themes. We propose to continue our efforts on methodologic issues in gene characterization, with the following aims: (i) Development of a unified likelihood framework for family-based association studies, including creation of a likelihood for segregation, linkage, direct association, and linkage disequilibrium for candidate genes, non-random ascertainment of families, censored age-at-onset phenotypes, estimation of penetrance from ascertained families with only some members genotyped, robust tests and variance estimators allowing for residual familial aggregation, multistage sampling of pedigrees, uncertainties in reported phenotype information in genetic analyses, implementation of these methods in the research version of the Genetic Analysis Package (GAP-A), and investigation of the relative efficiency of alternative study designs; (ii) Approaches to modeling complex disease traits, including methods for analysis of multiple phenotypes (e.g., multiple sites of cancer, such as the HNPCC syndrome), toxicokinetic models for complex metabolic pathways involving multiple genes and environmental exposures, and penetrance models incorporating genomic instability; (iii) Methods for characterizing the phenotypic effect of highly polymorphic genes; (iv) Methods for allowing for population stratification in case-control and cohort gene-association studies of unrelated individuals. To illustrate our methodologic work, we will draw examples from the CFRBCCS and other studies the genetic epidemiology of cancer in the USC Department of Preventive Medicine.