Towards an Internet of Science

Area of Science:

• Bioinformatics and computational biology.
• Data-driven scientific research.

Background:

• Big data and complex analysis workflows are prevalent in data-driven science, particularly bioinformatics.
• Numerous computational tools exist for data analysis, leading to duplicated effort (e.g., >50 read mapping tools).
• Current workflow management systems often integrate largely untested tools, yielding unpredictable results.

Purpose of the Study:

• To address the challenges of duplicated effort and tool reliability in scientific data analysis.
• To propose a new system for managing computational pipelines that is cloud-ready.
• To facilitate development and analysis in both public and private cloud environments.

Main Methods:

• The study identifies limitations in existing workflow management systems regarding cloud transition.
• It proposes a novel system architecture designed for cloud integration.
• Key features include overcoming duplicated efforts, introducing rigorous testing, and enabling cloud-based development and analysis.

Main Results:

• Current systems are not prepared for the shift towards cloud-based data analysis.
• A significant amount of duplicated effort exists in the development of computational tools.
• The reliability of integrated tools in current pipelines is often uncertain.

Conclusions:

• A new, cloud-ready workflow management system is needed to address current inefficiencies and reliability issues.
• The proposed system aims to reduce duplicated efforts, enhance tool testing, and support cloud-based scientific analysis.
• Future development should focus on creating robust, interactive, and cloud-integrated data analysis pipelines.