Development of Statistical Approaches Allowing for Genetic Covariates

描述(由申请人提供)： 为了确定一个复杂性状的遗传基础，有必要使用考虑到该性状背后的多个遗传成分的共同影响的方法。然而，这不可能使用通常的分离分析，这通常是有效的，只有当一个家庭中的性状变异主要是由于一个突变分离在一个假定的座位。针对这一点，我们建议在分离分析模型中加入针对突变携带者状态进行调整的遗传协变量，该分离分析模型考虑了复杂性状的遗传复杂性和异质性。包含遗传协变量的分离分析模型将比通常的分离分析模型更准确地模拟复杂的遗传效应。最近，我们在Li-Fraumeni综合征家系的分离分析中使用了一个独立的P53突变状态的遗传协变量。这项研究将于8月发表在《癌症研究》杂志上。然而，在该研究中使用独立的遗传协变量并没有充分考虑遗传突变分布中的家庭内相关性。当突变基因只对一个家庭中的一些亲属可用时，这个问题可能会更加复杂和严重。这项提案的中心主题是开发统计方法，允许相关的遗传协变量。在分离分析模型中，开发相关的遗传协变量来解释遗传突变分布中的家庭内相关性是一种新颖和可取的方法。我们建议使用基于模拟的方法来定量确定相关遗传协变量的影响。本项目将使用两种基于最大似然法的复杂分离分析模型作为年龄相关的风险模型。新开发的统计学方法将被用来确定与BRCA1/2突变相关的乳腺癌的遗传基础。癌症遗传网络的区域中心德克萨斯癌症遗传联合会将提供乳腺癌数据。

DESCRIPTION (provided by applicant): To determine the genetic basis of a complex trait, it is necessary to use methods that take account of the joint effects of multiple genetic components underlying the trait. However, this is not possible using a usual segregation analysis, which is often efficient only when the variation of the trait in a family is largely due to a mutation segregating at a single putative locus. In response to this, we propose to incorporate genetic covariates adjusted for mutation carrier status in segregation analysis models that account for the genetic complexity and heterogeneity of a complex trait. Segregation analysis models that include genetic covariates will be more accurate for modeling complex genetic effects than the usual segregation analysis models. Recently, we used an independent genetic covariate for p53 mutation status in the segregation analyses of families with Li-Fraumeni syndrome. This study will be published in Cancer Research in August. However, the use of independent genetic covariates in that study did not take full account of intrafamilial correlation in hereditary mutation distributions. This problem could be more complicated and serious when mutation genotypes are only available for some relatives in a family. The central theme of this proposal is to develop statistical approaches that allow for dependent genetic covariates. It is novel and desirable to develop dependent genetic covariates that account for intrafamilial correlation in hereditary mutation distributions in segregation analysis models. We propose to use simulation-based approaches to quantitatively determine the effects of dependent genetic covariates. Two types of complex segregation analysis models by maximum likelihood will be used as age-specific risk models in this project. The newly developed statistical approaches will be used to determine the genetic basis of breast cancer in association with mutations in BRCA1/2. The Texas Cancer Genetic Consortium, a Cancer Genetic Network regional center, will provide the breast cancer data.