Sequential pattern mining for discovering gene interactions and their contextual information from biomedical texts.

Area of Science:

• Bioinformatics
• Computational Biology
• Text Mining

Background:

• Extracting gene interactions from large biological text collections is challenging for researchers.
• Existing Natural Language Processing (NLP) methods are often rule-based, time-consuming, or produce non-interpretable machine learning outcomes.

Purpose of the Study:

• To develop an original symbolic method for automatically extracting gene interactions and their semantic characterizations from biological texts.
• To overcome limitations of existing NLP approaches by providing understandable and comprehensive gene interaction data.

Main Methods:

• Hybridization of data mining and natural language processing (NLP).
• Development of a symbolic method for pattern production.
• Utilizing text collections as training corpora.

Main Results:

• Successfully detected gene interactions and associated semantic information (modalities, contexts, types).
• Achieved results comparable to state-of-the-art methods, outperforming them in AIMed gene interaction detection.
• Demonstrated the method's effectiveness in discovering interactions and their characterizations.

Conclusions:

• The developed approach effectively discovers gene interactions and their associated semantics.
• Few existing methods automatically extract both interactions and semantic information.
• Extracted gene interaction data from PubMed is accessible via a web interface, and the software is publicly available.