Model-driven user interfaces for bioinformatics data resources: regenerating the wheel as an alternative to...

Area of Science:

• Bioinformatics
• Software Engineering
• Data Management

Background:

• Bioinformatics data repositories require diverse interfaces (web, desktop, command line) for data access.
• Existing interfaces often have duplicated functionality, leading to repeated development efforts and inefficient resource allocation.
• Limited developer resources in public laboratories necessitate streamlined user interface development.

Purpose of the Study:

• To propose a model-driven approach for the automated generation of user interfaces for bioinformatics data repositories.
• To reduce the time domain experts spend on repetitive user interface development tasks.
• To enable better resource allocation towards specialized tasks like data integration and analysis.

Main Methods:

• Developed a system named Pierre, inspired by the Object Management Group's Model Driven Architecture (MDA).
• The system partially generates user interfaces based on models describing service capabilities.
• Validated the approach through use cases in life sciences, including proteomics and data warehouse applications.

Main Results:

• Pierre successfully generates various interface types: web, stand-alone applications, text-menu, and command line.
• The system supports common database access requirements across different repository types (relational, XML, object databases).
• Demonstrated applicability to real-world bioinformatics resources like PEDRoDB and e-Fungi.

Conclusions:

• Model-Driven Architectures (MDAs) facilitate rapid development of repository interfaces in bioinformatics.
• The Pierre MDA effectively supports diverse database access needs with auto-generated interfaces.
• This approach enhances efficiency and flexibility in managing bioinformatics data resources.