SimText: a text mining framework for interactive analysis and visualization of similarities among biomedical entities

Area of Science:

• Bioinformatics
• Computational Biology
• Text Mining

Background:

• Biomedical literature exploration, particularly assessing associations between entities like genes and diseases, is complex and time-consuming.
• Existing methods often lack systematic workflows for analyzing entity similarities based on textual data.

Purpose of the Study:

• To introduce SimText, a user-friendly toolset designed for systematic analysis of similarities among biomedical entities.
• To provide customizable workflows for text collection, entity extraction, and interactive data visualization.

Main Methods:

• SimText utilizes text mining approaches for word extraction from PubMed literature.
• It employs unsupervised learning techniques for interactive data analysis and visualization.
• The toolset is implemented as open-source R software and integrated into the Galaxy platform.

Main Results:

• SimText enables efficient text collection and entity-specific word extraction from PubMed.
• The tool facilitates interactive analysis and visualization of entity similarities.
• Customizable workflows support systematic exploration of biomedical literature.

Conclusions:

• SimText offers a powerful and accessible solution for researchers to analyze biomedical entity similarities.
• The integration with Galaxy and availability of training materials enhance its usability.
• This toolset can significantly improve the efficiency of literature-based biomedical research.