SOLAR-Eclipse Computational Tools for Imaging Genetics

描述(申请人提供)：该申请将为新兴的成像遗传学领域提供急需的分析方法。我们的重点是创建日食成像遗传学工具，用于经典的遗传和表观遗传流行病学分析，如遗传率、多效性、数量性状基因座(QTL)和全基因组关联(GWAS)、基因表达和甲基化分析，使用来自结构和功能脑成像(AIM 1)的性状。我们还将开发同时满足遗传和成像要求的多项测试的智能校正(AIM 2)。太阳-日食工具将进行优化，以使用体素脑功能和结构成像特征以及高通量现代分子遗传数据。拟议的工具将是灵活的，既可以容纳结构化的家庭流行病学样本，也可以容纳非结构化的流行病学样本。这些工具将作为一个独立的应用程序发布，并集成到现有的神经科学eScience网络中。在单一分析包中实现这一目标将极大地增强和加快对影响大脑神经解剖和功能特征的基因的搜索，并提供全面的工具来说明从基因到大脑结构/功能的途径。在开发过程中，我们将使用由NIH资助的三个成像遗传学项目、大脑结构的遗传项目、人类连接项目和NicotineBrain生成的数据来测试和优化这些工具(AIM 3)。这些工具、源代码和书面分析策略将通过神经成像信息学工具和资源信息中心(NITRC.org)分发。这个应用程序的PI是一位由NIBIB资助的K奖获得者，他致力于优化成像遗传学研究的标准遗传分析方法。这项工作在过去四年中产生了25份同行评议的出版物。这些经验将被用来创建一个新的工具库，名为SOLAR-日食，它将基于NIH资助的统计遗传方法库。我们的初步数据显示，当针对成像特征进行优化时，这些工具可以实现100-200倍的性能改进(相对于标准方法)，从而能够在多节点Linux集群上进行全面的(103个受试者的105个特征)体素单变量和双变量遗传分析。这笔赠款的合作者结合了成像和遗传学方面的专家，他们展示了在这两个领域开发工具的经验，并开发了流行的分析工具，如SOLAR、FSL、SPM、BrainMap、Talairach Daemon、JIST、Mango等。 公共卫生相关性：融合了现代科学中的两个重要研究方向，遗传学和神经成像，导致了一个新领域的出现，称为成像遗传学。这一领域结合了现代统计遗传学方法和用高维神经成像方式进行的定量表型。这门学科试图通过表征导致风险的基因来了解神经和精神疾病的生物学基础。 治疗这些疾病。然而，标准的成像工具无法处理大规模的遗传学数据，而标准的遗传学工具又无法适应大尺寸和大格式的成像特征。因此，迫切需要优化的软件工具来执行单变量和多变量成像遗传学分析，同时为多个测试提供实用的校正策略。

DESCRIPTION (provided by applicant): This application will provide urgently needed analysis methods to the emerging field of imaging genetics. Our focus is to create SOLAR-Eclipse imaging genetics tools for classical genetic and epigenetic epidemiological analyses such as heritability, pleiotropy, quantitative trait loci (QTL) and genome-wide association (GWAS), gene expression, and methylation analyses using traits derived from structural and functional brain imaging (AIM 1). We will also develop intelligent correction for multiple testing that meets both genetic and imaging requirements (AIM 2). The SOLAR-Eclipse tools will be optimized to use voxel-wise brain functional and structural imaging traits and high throughput modern molecular genetic data. The proposed tools will be flexible to accommodate both structured family-based and unstructured epidemiological samples. These tools will be released as a standalone application and integrated into the existing neuroscience eScience networks. Achieving this goal in a single analysis package will greatly enhance and speed up the search for genes that influence brain's neuroanatomic and functional traits and provide comprehensive tools to illustrate pathways from genes to brain structure/function. During the development process we will test and optimize these tools (AIM 3) using the data generated by three NIH-funded imaging genetics projects, the Genetic of Brain Structure, the Human Connectom project, and the NicotineBrain. The tools, source codes and documented analyses strategies will be distributed via the Neuroimaging Informatics Tool and Resources Clearing house (NITRC.org). The PI for this application is a NIBIB-funded K-awardee who has worked on optimizing the standard genetic analysis methods for imaging genetics research. This work generated twenty-five peer- reviewed publications in the last four years. These experiences will be used to create a new library of tools termed, SOLAR-Eclipse, which will be based on the NIH-funded SOLAR library of statistical genetic methods. Our preliminary data shows that when optimized for imaging traits these tools can achieve 100-200 fold performance improvement (over standard methods) enabling full scale (~105 traits in 103 subjects) voxel-wise univariate and bivariate genetic analyses on a multi-node Linux cluster. The collaborators on this grant combine experts in imaging and genetics who have demonstrated their experience in developing tools in both domains and developed popular analyses tools such as SOLAR, FSL, SPM, BrainMap, Talairach Daemon, JIST, Mango and others. PUBLIC HEALTH RELEVANCE: Merging two important research directions in modern science, genetics and neuroimaging, led to the emergence of a new field, termed imaging genetics. This field combines modern statistical genetic methods with quantitative phenotyping performed with high dimensional neuroimaging modalities. This discipline seeks to understand the biological basis of neurological and psychiatric illnesses by characterizing genes that contribute to the risk for these disorders. However, standard imaging tools are unable to deal with large-scale genetics data and standard genetics tools, in turn, are unable to accommodate large size and format of the imaging traits. Thus, there is in immediate need for software tools optimized for performing univariate and multivariate imaging genetics analyses while providing practical correction strategies for multiple testing.