User-ready tools and scalable workflows for INCLUDE datasets in the cloud: advancing brain imaging data management and analytics

项目总结 随着医疗数据集和数据存储库的快速增长，对技术基础设施的需求非常迫切 使数据共享和可用性成为可能。云生态系统提供了计算效率高的资源 大规模数据处理和分析。虽然基于云的平台已经为分子和 在临床数据处理方面，目前还没有单一的端到端平台来处理生物医学影像数据。 此外，没有标准化的系统来实现分子、临床和 成像数据管道。我们建议(目标1)将现有的开源神经成像软件集成到 基于云的集中式脑成像数据平台(飞轮)。这将提供可扩展、自动化和 可公开访问的工具，用于处理和分析来自标准成像模式的数据，并提供简化的 用于标准化和隐私保护(去身份)的管道，用于提交给长期存储库。 此外，我们将(目标2)建立数据表示、本体和工作流以实现互操作性 在目前不兼容的分子和成像系统(CAVATICA和飞轮)之间。这将允许 在脑病理和发育的科学研究中使用丰富的多模式数据集。现在 建议利用软件开发人员和生物医学的跨学科团队之间的合作 分析师和研究人员，以及他们在计算机科学、云计算和 神经成像和分子研究。总体而言，该项目将提供基础设施和方法来 支持神经科学和生物信息学社区的大规模数据托管和管理， 将支持包括多机构研究和临床试验在内的科学努力。

PROJECT SUMMARY With rapidly growing medical datasets and data repositories, there is a critical need for technical infrastructures to make data sharing and usability feasible. Cloud ecosystems provide computationally efficient resources for large-scale data handling and analytics. While cloud-based platforms have been developed for molecular and clinical data processing, there is no single end-to-end platform for processing biomedical imaging data. Moreover, there is no standardized system to enable interoperable work between molecular, clinical, and imaging data pipelines. We propose to (Aim 1) integrate existing open-source neuroimaging software into a centralized, cloud-based platform for brain imaging data (Flywheel). This will provide scalable, automated, and publicly accessible tools to process and analyze data from standard imaging modalities, and supply streamlined pipelines for standardization and privacy-protection (de-identification) for submission to long-term repositories. In addition, we will (Aim 2) establish data representation, ontology, and workflows to allow for interoperability between currently incompatible molecular and imaging systems (CAVATICA and Flywheel). This will allow the use of rich multi-modal datasets in scientific research on brain pathology and development. The present proposal leverages collaboration between inter-disciplinary teams of software developers and biomedical analysts and researchers, and their complementary expertise in computer science, cloud computing, and neuroimaging and molecular research. Overall, this project will provide the infrastructure and methods to support large-scale data hosting and management for the neuroscience and bioinformatics communities, which will support scientific efforts including multi-institutional research and clinical trials.