"Statistical Methods for Genetic Epidemiology"

描述（由申请人提供）：对有或无疾病个体的大样本的高维单核苷酸（SNP）数据的可用性为遗传流行病学家提供了前所未有的机会，但也带来了许多统计挑战。我们实现这些机会的能力取决于发展和应用健全和创新的统计方法。特别需要的领域包括处理多个假设检验问题（多个比较问题）的方法，以及组合来自多个来源的数据以获得更清晰推论的方法（数据综合问题）。这项研究的目标是开发新的或改进的方法来解决这些问题，并将其应用于乳腺癌、卵巢癌和前列腺癌的数据。为了实现这些目标，研究人员将在过去15年的工作基础上继续努力。具体来说，1988年，美国国家癌症研究所（NCI）向首席研究员颁发了杰出研究者奖（OIG），以表彰他在流行病学研究中开发和应用了新的、改进的统计方法。1995年，这项资助被续期，直到2001年NCI终止该项目。该研究由R01 CA94069资助，时间为2002年1月至2006年12月。这项竞争性的续期申请要求资金继续这项工作，并将其应用于日益重要的新领域。具体目的有四个方面：1)评估病例对照全基因组关联（GWA）研究中控制混杂和尾计数变异性的新方法和现有方法；2)评估已确定疾病相关性基因中意义未知的错义突变的相关风险；3)结合多个独立的数据来源来评估基因群的联合致癌效应；4)整合肿瘤和患者特征的数据，以改进对表型特异性病因的评估。我们将通过模拟来评估这些新方法，并将它们应用于乳腺癌、卵巢癌和前列腺癌的现有数据，以及来自乳腺癌和前列腺癌GWA研究的新数据。现有数据要么由研究人员和该项目的顾问收集，要么作为研究人员参与的三个合作项目的一部分收集：乳腺癌家庭登记处（Breast- cfr）、卵巢癌协会联合会（OCAC）和国际前列腺癌遗传学联合会（ICPCG）。总之，我们需要健全、可靠的方法来分析来自患有或不患有特定类型癌症的个体的大量新出现的基因数据。这项研究的目标是开发这样的方法，从而帮助产生对癌症预防和治疗有用的新知识。

DESCRIPTION (provided by applicant): The availability of high-dimensional single-nucleotide (SNP) data for large samples of individuals with and without disease presents unprecedented opportunities for genetic epidemiologists, but also many statistical challenges. Our ability to bring these opportunities to fruition depends on the development and application of sound and innovative statistical methods. Areas of particularly great need include methods for dealing with multiple hypothesis-testing problems (multiple comparison issues), and methods for combining data from multiple sources to achieve sharper inferences (data synthesis issues). The goals of this research are to develop new or improved methods for addressing these issues, and to apply them to data on cancers of the breast, ovary and prostate. To accomplish these goals, the investigators will build on more than 15 years of previous work. Specifically, in 1988 the National Cancer Institute (NCI) awarded an Outstanding Investigator Grant (OIG) to the Principle Investigator for the development and application of new and improved statistical methods for use in epidemiological research. In 1995 this grant was renewed until NCI terminated the program in 2001. The research was then funded by R01 CA94069 for the period January 2002 to December 2006. This competing renewal application requests funds to continue this work and to apply it to new areas of emerging importance. The specific aims are fourfold: 1) to evaluate new and existing methods for controlling confounding and tail count variability in case-control genome-wide association (GWA) studies; 2) to evaluate risks associated with missense mutations of unknown significance in genes of established disease relevance; 3) to combine multiple independent sources of data in assessing the joint carcinogenic effects of groups of genes; and 4) to integrate data on tumor and patient characteristics for improved assessment of phenotype-specific etiology. We will evaluate the new methods by simulations, and we will illustrate them by application to existing data on cancers of the breast, ovary and prostate, and emerging data from GWA studies of breast and prostate cancer. The existing data have been collected either by the investigators and the consultants to this project, or as part of three collaborations in which the investigators participate: the Breast Cancer Family Registry (Breast-CFR), the Ovarian Cancer Association Consortium (OCAC), and the International Consortium for Prostate Cancer Genetics (ICPCG). In summary, we need sound, reliable methods for analyzing the vast amounts of emerging genetic data from individuals with and without a given type of cancer. The goals of this research are to develop such methods, and thereby to help generate new knowledge useful for cancer prevention and treatment.