Our long-term objective has been to develop statistical methods for fitting models for complex traits to family data in order to test hypotheses about genetic effects and gene-environment interactions and to localize genes. We have pursued a variety of approaches, including parametric models based on an extension of the proportional hazards model to family data to incorporate major genes and polygenes; MCMC and GEE methods for fitting complex genetic likelihoods; Bayesian methods for models with many parameters; and investigation of the bias and relative efficiency of various study designs In this cycle, we propose to split our activities into two separate grants: in this competing continuation application, we focus on the design and analysis of family studies for testing candidate gene associations and residual familial aggregation; a new proposal will continue the research on computational methods. Specifically, in this application, we propose to; 1. Investigate study designs and analysis methods that allow for efficient and unbiased estimation of population parameters (e.g. penetrance and population allele frequency); 2. Develop models for """"""""complex diseases"""""""" (including GCS and gag interactions, age effects, etc.); 3. Develop conditional and marginal model approaches to allow for residual familial aggregation in major gene models for multivariate survival data; and 4. Explore extensions of these designs for gene mapping studies aimed at finding genes that interact with environmental agents.
Showing the most recent 10 out of 63 publications
