生物資訊分析自動化流程管理系統

Abstract

整合分散且異質的生物資訊資料庫及軟體資源持續的吸引研究人員的投入，而最近幾年的網路服務(Web Service)開啟了生物資訊整合系統的新浪潮。由於網路服務的標準都還在制定中，對於生物資訊複雜的環境能帶來多少裨益還是個未知的問題。在這篇論文中我們以網路服務協調為核心，提出一個生物資訊服務協調平台-BioOrch,去描繪出以網路服務為基礎的整合系統應該要包含哪些重要的設計考量。BioOrch將流程邏輯中的活動分離出來，讓它可以更簡單且更容易延伸。 此外我們將活動定義為自我滿足的執行單元，使的活動可以在不同的工作流程管理系統之間交換執行。我們展示了這樣的模式可以大大的簡化流程管理系統的設計與實作，並讓不同組織的人可以互相合作。

Integration of distributed, heterogeneous biological databases and software resources continues to attract research interests, and the recent Web services movement has also generated a new wave of development of integrated bioinformatics systems. As Web service standards are still evolving, to what degree bioinformatics integration can be achieved using Web services remains a question to be answered. In this thesis, we focus on Web service orchestration aspect, and present a service orchestration framework for bioinformatics, called BioOrch, to illustrate important design considerations when building Web services-based integrated systems. In particular, BioOrch separates flow logic from activities in a workflow, making each activity simpler and more extensible. Furthermore, as activities are self-describing, they can be exchanged among different workflow managers. We show that this model can greatly simplify the design and implementation of cross-organizational workflow systems involving people from multiple sites.