A Robust, Secure Framework to Effortlessly Bind Distributed Databases and Analysis Tools into Tightly Integrated Translational Drug Discovery Computational Platforms

项目摘要 Collaborative Drug Discovery，Inc. (CDD)建议为药物发现开发云工作空间- a 新的信息学框架，将使从事药物发现和翻译的科学家能够努力- 更少、更健壮、更安全地集成分布在多个 系统和供应商整合到高效、定制的计算工作流程中。我们的创新技术- ogy将解决一个阻碍药物发现和翻译工作的关键问题：该领域的科学家通常会 通常需要将来自多个来源的化学和生物数据联合收割机，通过多个软件运行它们 专门用于不同类型的分析和可视化的软件包，然后理想地存储 分析及其基础实验数据。今天，这种类型的整合是困难的， 实现成本高，而且通常很脆弱，这对（i）利用快速增长的 高质量的公共访问数据储存库的数量，以及㈡评价有前途的新分析工具， 战略布局单片平台通过将所有东西集中在一个屋檐下来解决这个问题， 它们非常昂贵，并且限制了灵活性：没有一个平台可以提供所有功能。另一个- 本机方法-将离散资源串在一起-在桌面计算时代发展， 无法很好地转化为现代基于云的工作流程，尤其是在执行 需要组合分布在远程系统中的大型数据集的计算密集型操作。 云工作空间（CW）旨在联合收割机结合这两个极端的优点，避免这两个极端的缺点。CW 从本质上讲，将允许用户轻松创建自己的个性化云托管解决方案， 独特的要求和工作流程。我们的方法提供了性能，鲁棒性和易用性， 整体软件解决方案，但没有相关的兼容性和供应商锁定。它提供了灵活性 和开放性，但没有相关的整合挑战和脆弱性， 它推进了流水线方法，以包含基于云的模型和分布式数据， 资源，而不会影响性能或安全性。在第一阶段，我们证明了我们可以稳健地， 有效地同步生物和化学数据（仅传输新的或修改的数据，同时保留 化学标识符的正确关联）在CW容器环境和远程数据库之间， 这对我们的概念来说是一个具有挑战性但必不可少的先决条件。在第二阶段，我们将完成 CW并通过多个实际应用程序和软件应用程序证明其有效性 合作伙伴和测试版客户最终用户。该技术的市场范围从学术界到小型和 从中型企业到大型制药企业。

PROJECT SUMMARY Collaborative Drug Discovery, Inc. (CDD) proposes to develop Cloud Workspaces for Drug Discovery – a novel informatics framework that will enable scientists engaged in drug discovery and translation to effort- lessly, robustly, and securely integrate disparate databases and computational tools distributed across multiple systems and vendors into highly-efficient, custom-tailored computational workflows. Our innovative technol- ogy will solve a critical problem that hinders drug discovery and translation efforts: scientists in this field typi- cally need to combine chemical and biological data from several sources, run them through multiple software packages that specialize in different types of analyses and visualization, then ideally store the results of the analyses together with their underlying experimental data. Today, this type of integration is difficult and expensive to accomplish and typically fragile, creating a large barrier to (i) exploiting the rapidly increasing number of high-quality public-access data repositories and (ii) evaluating promising new analytical tools and strategies. Monolithic platforms offer to solve this problem by bringing everything together under one roof, but they are extremely expensive and they limit flexibility: no single platform can offer every capability. The alter- native approach – stringing together discrete resources – evolved during the era of desktop computing and does not translate well to modern cloud-based workflows and in particular to the challenges of performing computationally intensive operations that require combining large datasets distributed across remote systems. Cloud Workspaces (CW) aims to combine the strengths and avoid the weaknesses of these two extremes. CW will in essence allow users to easily create their own individualized cloud-hosted solutions tailored to their unique requirements and workflows. Our approach offers the performance, robustness, and ease of use of a monolithic software solution, but without the associated inflexibility and vendor lock in. It offers the flexibility and openness of combining discrete resources, but without the associated integration challenges and fragility, and it advances the pipelining approach to embrace cloud-based models and to encompass distributed data resources without compromising performance or security. In Phase 1 we proved that we could robustly and efficiently synchronize biological and chemical data (transferring only new or modified data while retaining correct association of chemical identifiers) between the CW container environment and remote databases, which was a challenging but essential prerequisite for our concept. In Phase 2 we will complete development of CW and demonstrate its effectiveness with multiple real-world applications together with software application partners and beta customer end users. The market for the technology ranges from academics to small and medium size companies to the large pharmaceutical firms.