SEI(BIO)+II: A Collaborative Scientific Workflow Environment for Accelerating Genome-Scale Biological Research

该项目将开发一个系统，该系统集成了基于web的实验工作空间环境和灵活的科学工作流自动化系统，支持快速原型和用户透明的分布式计算。实验工作空间网络界面将为生物学家提供一个可扩展的虚拟实验室，用于定义计算协议，执行基于这些协议的“实验”，在其来源的背景下可视化数据，以及管理项目。开普勒科学工作流程系统将被扩展到在可用的分布式资源上自动执行这些实验，并将记录工作流程和数据来源。生物信息学专家和软件工程师将使用该系统开发新的实验模板，并与实验工作空间的用户共享这些模板。该系统将支持能够有效利用这两个接口的用户，并将促进生物学家、生物信息学专家和软件开发人员之间的高效协作。该系统的原型将集成并加速研究项目的计算和实验组件，采用顺序染色质免疫沉淀，然后进行DNA微阵列分析（ChIP-chip），以识别直接转录因子靶标。自动化ChIP-chip工作流程中的IT和CS挑战是典型的困扰复杂的基因组规模分析(例如，同时处理多个，替代或相关的分析模块，管理数千个工作流程运行和结果数据集，记录工作流程，数据和参数依赖关系；实现这一系统将需要创新的方法来管理嵌套的科学数据集合，将函数式编程和科学工作流的流处理方法结合起来，并推广数据类型和集成方法，以支持不同组织开发的工作流组件的互操作性。该提案汇集了在科学工作流建模，设计和自动化方面的专业知识的研究人员；科学数据管理；芯片研究和数据分析；开发用于科学研究的协作计算环境，这将共同提高学生对研究的参与度。

This project will develop a system that integrates a web-based experiment workspace environment with a flexible scientific workflow automation system supporting rapid prototyping and user-transparent distributed computing. The experiment-workspace web interface will provide biologists an extensible virtual laboratory for defining computational protocols, executing "experiments" based on those protocols, visualizing data in the context of their provenance, and managing projects. The KEPLER scientific workflow system will be extended to automate the enactment of these experiments on available distributed resources and will record workflow and data provenance. Bioinformatics specialists and software engineers will develop new experiment templates using the system and will share these templates with experiment-workspace users. The system will support users who can make effective use of both interfaces and will facilitate productive collaboration between biologists, bioinformatics specialists, and software developers. A prototype of this system will integrate and accelerate computational and experimental components of research projects employing a sequential chromatin immunoprecipitation followed by DNA microarray analysis (ChIP-chip) for identifying direct transcription factor targets. The IT and CS challenges in automating ChIP-chip workflows are typical of those plaguing complex, genome-scale analyses (e.g., dealing with multiple, alternate or related analysis modules simultaneously; managing thousands of workflow runs and resulting datasets; recording workflow, data and parameter dependencies; etc.) Implementing this system will require innovative approaches for managing nested collections of scientific data, combining functional programming and stream-processing methodologies for scientific workflows, and generalizing data-typing and integration approaches for supporting interoperability of workflow components developed by different organizations.This proposal brings together investigators recognized for their expertise in scientific workflow modeling, design, and automation; scientific data management; ChIP-chip research and data analysis; and development of collaborative computational environments for scientific research, which together will enhance student engagement in the research.