Filling the gaps between tools and users: a tool comparator, using protein-protein interaction as an example

Area of Science:

• Bioinformatics
• Computational Biology
• Natural Language Processing

Background:

• Text mining (TM) and Natural Language Processing (NLP) tools are nearing practical performance levels.
• Current NLP/TM systems lack reusability and interoperability, hindering deployment, especially in diverse BioNLP applications.
• Existing frameworks like UIMA offer partial solutions, but require further development for true toolkit interoperability.

Purpose of the Study:

• To describe a novel UIMA-based environment designed to enhance the reusability and interoperability of NLP/TM tools.
• To address the challenges of deploying NLP/TM technology in specialized domains like BioNLP.
• To demonstrate the feasibility of the developed environment through a practical application.

Main Methods:

• Development of a UIMA-based environment incorporating sharable type systems and a developer-friendly interface.
• Integration of features for tool comparison, simplified I/O, and visualization.
• Application of the environment to build a protein-protein interaction (PPI) extraction system.

Main Results:

• The developed environment facilitates the integration and reuse of NLP/TM tools.
• The protein-protein interaction (PPI) extraction system demonstrated the practical feasibility of the environment.
• The system showcases improved efficiency and reduced development time for specialized BioNLP tasks.

Conclusions:

• The UIMA-based environment significantly improves the reusability and interoperability of NLP/TM tools for BioNLP.
• This approach addresses the diverse requirements of biologists by providing a flexible and integrated platform.
• The developed environment is crucial for advancing the practical application of text mining in biological research.