A major emphasis will be placed on the development and evaluation of test statistics for candidate gene association with disease risk, taking explicit account of the ages at onset of affected individuals. These test statistics are based on the familiar transmission/disequilibrium test and the related sibling-TDT, making use of prior information of the penetrance function, it is known. These statistics will be generalized to accommodate a general pedigree structure, haplotypes, and gene- environment interactions. A large amount of genetic data is and more will be generated using array technologies. This leads to statistical problems with a high-dimensional predictive space. Methods that were developed recently in computational statistics will be adapted for assessing gene and disease association. Specifically, logic regression, using a Boolean combination of """"""""and"""""""" and """"""""or"""""""" with logic statements of predictors, will be developed in order to enhance the interpretability of the models. Multiple testing issues will be explored. Methods for characterizing the relationship the relationship between genes and disease risk (hazard rate) and for assessing gene-environment interactions will be developed and evaluated. Methods for gene characterization will also be investigated in the situation that the disease prevalence is relatively common and that the ages at onset of subjects are correlated even after adjusting for observed genetic and environment risk factors. Collectively, the proposed research has the potential to enhance the scientific and public health knowledge through better methods for design, conduct, and analysis of genetic epidemiological studies.
Showing the most recent 10 out of 319 publications
