SAMPLING MODELS AND METHODS FOR COMPLEX GENETIC DISEASES

点击翻译按钮获取中文摘要

DESCRIPTION (Adapted from investigator's abstract): This is a competitive renewal application. The original proposal developed statistical methods for addressing sampling issues arising in human genetics research. The current proposal extends this work by developing methods to handle sampling differences across genetic linkage studies. Discovery of specific genetic influences on the development of complex disorders in humans can facilitate early diagnosis and prevention, and broaden understanding of the contributions of both nature and nurture to the development of illness. In recent years, increased interest in complex disorders has fueled rapid development of complex investigative techniques, which in principle makes it possible to map genes of even small to moderate effect. However, a dearth of methods for what is sometimes called 'meta-analysis,' or the mathematically rigorous evaluation of the overall statistical evidence based on multiple sets of genetic data, has made it extremely difficult in practice to interpret the aggregate results of even the most sophisticated data analyses. The aim of this proposal is to develop and evaluate mathematically rigorous methods for representing the strength of the genetic evidence based on multiple sets of data, where the data are inherently heterogeneous. We propose a novel approach to this problem, which we call sequential Bayesian. This approach has conceptual and computational advantages over existing alternatives, and offers a fully general way to handle differences across data sets. The new approach will be thoroughly developed and evaluated by comparison with existing methods, based on both analytic work and simulations studies. The objective will be to produce a comprehensive set of guidelines for investigators interested in analyzing multiple sets of genetic data. The project will also serve as a critical adjunct to ongoing studies of autistic disorder and panic disorder, for which the question of how to interpret accumulated genetic linkage evidence, is increasingly urgent.