Computational Efficient Statistical Tools for Analyzing Substance Dependence Sequencing Data

项目摘要 随着下一代测序技术的进步，测序研究已经变得越来越重要。 用于物质依赖（SD）研究。这些研究产生了大量的测序数据 并使研究人员能够全面调查SD中遗传变异的深层目录的作用。 尽管正在进行的测序研究在揭示新的变异方面有很大的希望， SD，高维数据，低频率变体，复杂SD病因和异质SD 表型产生巨大的分析和计算挑战。开发强大的 方法和计算效率高的软件将解决SD测序数据分析中的挑战 并增强我们识别新的SD相关变体的能力。该应用程序的目标是开发新的 设计和分析基于群体和基于家族的测序数据的方法和软件 单个或多个表型，并在合作研究中使用它们来研究遗传变异， 与SD相关的基因-基因/基因-环境（G-G/G-E）相互作用。根据初步模拟 结果，我们的中心假设是，所提出的方法比现有的计算效率更高 方法，并获得更强大和强大的性能为各种类型的表型。计划 具体目标是：1）开发一种新的非参数方法用于测序数据的设计和分析 2）发展一种联合U方法用于高维G-G/G-E相互作用 3）建立了一种基于家系的SD测序的家系相似性U方法 数据分析，解释人群分层和家庭中丰富的罕见变异;以及4）促进 通过软件开发和协作使用新方法。拟议的研究将是 由一位早期的新研究者（NIDA K 01获奖者）发起，他组建了一个科学家团队， 在统计遗传学、生物信息学/软件开发、SD流行病学、行为遗传学、 和临床精神病学。该项目的成功完成将解决若干重要的统计和 正在进行的测序研究中的计算差距，并推进方法和软件开发 用于SD测序数据分析。新方法和软件在大规模SD中的应用 测序数据集还有望发现新的SD相关变体和G-G/G-E 相互作用，这将最终导致更好地了解SD病因，从而产生潜在的好处 用于SD预防和治疗。

Project Summary With advancements in next-generation sequencing technologies, sequencing studies has become increasingly used in substance dependence (SD) research. These studies generate a massive amount of sequencing data and allow researchers to comprehensively investigate the role of a deep catalog of genetic variants in SD. Although the ongoing sequencing studies hold great promise for unraveling novel variants that contribute to SD, the high-dimensional data, low frequent variants, complex SD etiology, and heterogeneous SD phenotypes create tremendous analytic and computational challenges. Developing robust and powerful methods and computationally efficient software will address the challenges in SD sequencing data analysis and enhance our ability to identify new SD-related variants. The goals of this application are to develop new methods and software for designing and analyzing population-based and family-based sequencing data with single or multiple phenotypes, and to use them in collaborative research to investigate genetic variants and gene-gene/gene-environment (G-G/G-E) interactions associated with SD. Based on the preliminary simulation results, our central hypothesis is that the proposed methods are more computationally efficient than existing methods, and attain a more robust and powerful performance for various types of phenotypes. The planned specific aims are to: 1) develop a new non-parametric method for the design and analysis of sequencing data with one or multiple SD phenotypes; 2) develop a Joint-U method for high-dimensional G-G/G-E interaction analysis with SD sequencing data; 3) develop a family-similarity-U method for family-based SD sequencing data analysis, accounting for population stratification and rare variants enriched in families; and 4) facilitate the use of the new methods through software development and collaboration. The proposed research will be initiated by an early-stage new investigator (NIDA K01 awardee), who has assembled a team of scientists with expertise in statistical genetics, bioinformatics/software development, SD epidemiology, behavioral genetics, and clinical psychiatry. The successful completion of this project will address several important statistical and computational gaps in ongoing sequencing studies, and advance the methodology and software development for SD sequencing data analysis. The application of the new methods and software to large-scale SD sequencing datasets also holds promise for the discovery of new SD-associated variants and G-G/G-E interactions, which will ultimately lead to a better understanding of SD etiology, with resulting potential benefits for SD prevention and treatment.