PEDL+: protein-centered relation extraction from PubMed at your fingertip

Area of Science:

• Bioinformatics
• Computational Biology
• Natural Language Processing

Background:

• Relation extraction (RE) is crucial for biological database curation and pathway reconstruction.
• Existing RE systems often lack flexibility, being either non-programmable websites or fixed-function research code.
• Adaptable RE tools are needed for integration into complex omics data analysis pipelines.

Purpose of the Study:

• To present PEDL+, a user-friendly tool for extracting protein-protein and protein-chemical associations from PubMed.
• To provide a flexible and programmable solution for RE tasks in biological research.
• To facilitate the integration of RE into existing data analysis workflows.

Main Methods:

• PEDL+ utilizes state-of-the-art Natural Language Processing (NLP) technology.
• The tool incorporates adaptable ranking and filtering options for customized extraction.
• It is designed for easy integration into complex biological data analysis pipelines.

Main Results:

• PEDL+ was evaluated in two pathway curation projects.
• The tool demonstrated high utility, with 59% to 80% of its extractions found to be helpful.
• It offers a practical solution for extracting specific biological associations from large text corpora.

Conclusions:

• PEDL+ provides a valuable and user-friendly solution for relation extraction from biomedical literature.
• Its flexibility and integration capabilities make it suitable for diverse biological research applications.
• The tool enhances the efficiency of tasks like pathway reconstruction and omics data analysis.