The Cancer Genome Atlas Data Analysis Center

TCGA的首要目标是改变癌症治疗的做法，改善患者 通过癌症基因组学存活。一项关键交付成果是支持访问和使用复杂的多个 为下游研究提供维度基因组数据。我们建议运营一个GDAC-A中心， 开发分析管道所需的领导力、专业知识和基础设施 定义的综合分析和解释是为假设检验量身定制的， 翻译和临床研究人员。我们的团队由癌症生物学、基因组学和 生物信息学，在TCGA具有领导地位的记录。分析工具和管道结构为 基于我们丰富的TCGA经验，旨在以最佳方式实现其目标。这条管道将 使用GenePattern生物信息化工作流环境构建--灵活且模块化的体系结构 这是符合caBIG和caGRID的，在完善、强大和安全的IT基础架构中维护 在布罗德研究所，可以作为一条生产管道全天候运行。 利用这一成熟的资源，我们将实现以下具体目标。 目标1.我们将为所有输入和输出文件定义符合caBIG的数据格式。以进一步增强 标准化，我们建议在试点项目(级别)中定义的标准数据结构基础上增加两项内容 1-4)。0级将定义分析中使用的所有参考数据库的具体版本，5级将定义 捕捉纳入先验知识的疾病水平的发现。 目标2.我们将设计分析模块，以整合TCGA所有组件的数据，并 进行综合分析。结果将以符合caBIG标准的输出文件形式提交给DCC 附有人类可读的报告，其中包含文本摘要、表格和数字，格式为 不同学科的科学家都能理解，类似于出版物的结果部分。在……里面 此外，我们致力于管道的持续技术和分析改进，特别是 在支持向下一代测序平台过渡方面。 目标3.我们将在工业级生产中实现这种高通量分析管道 具有严格质量控制的模式，利用远大的基础设施支持和丰富的经验 在运行和维护高吞吐量计算流水线方面。

The overarching goal of TCGA is to change the practice of cancer nnedicine and improve patient survival through cancer genomics. A key deliverable is to enable access and use of complex multi- dimensional genomic data for downstream studies. We propose to operate a GDAC-A center with the leadership, expertise and infrastructure required to develop an analysis pipeline that will generate pre- defined integrative analyses and interpretations that are tailor-designed for hypothesis-testing by basic, translational and clinical investigators. Our team consists of experts in cancer biology, genomics and bioinformatics with a track record of leadership in TCGA. The analytical tools and pipeline structure are based on our extensive TCGA experiences and designed to optimally achieve its goals. This pipeline will be built using the GenePattern bioinformatic workflow environment - a flexible and modular architecture that is caBIG and caGRID compliant, maintained in the well-established, robust and secure IT infrastructure at the Broad Institute and can be operated 24/7 as a Production Pipeline. Leveraging this well-established resource, we will pursue the following specific aims. Aim 1. We will define caBIG compliant data format for all input and output files. To further enhance standardization, we propose two additions to the standard data structure defined in the Pilot Project (Levels 1-4). Level 0 will define specific versions of all reference databases used in the analyses and Level 5 will capture disease-level findings that incorporate prior knowledge. Aim 2. We will design analysis modules to consolidate data from all components of TCGA and to perform integrative analyses. Results will be submitted to DCC in caBIG compliant output files accompanied by human-readable reports containing text summaries, tables and figures in a format understandable to scientists of diverse disciplines, similar to the Results Section of a publication. In addition, we are committed to continuous technical and analytical improvement of the pipeline, particularly in supporting the transition to next-generation sequencing platforms. Aim 3. We will implement this high-throughput analysis pipeline in an industrial-level production mode with rigorous quality control, leveraging the Broad's infrastructural support and extensive experiences in running and maintaining high-throughput computational pipelines.