FLUTE: Fast and reliable knowledge retrieval from biomedical literature

Area of Science:

• Biomedical Informatics
• Computational Biology
• Bioinformatics

Background:

• Automated machine reading methods extract vast amounts of biomolecular interaction data from biomedical literature.
• A significant portion of these extracted interactions are inaccurate or irrelevant for specific computational modeling tasks.
• Efficient filtering methods are crucial for selecting high-quality data for systems biology.

Purpose of the Study:

• To develop and evaluate a tool for filtering and scoring biomolecular interactions extracted from biomedical literature.
• To improve the accuracy and relevance of data used in computational modeling.
• To accelerate the process of selecting reliable interaction data.

Main Methods:

• The FiLter for Understanding True Events (FLUTE) tool was developed to filter machine-extracted interactions.
• FLUTE utilizes public protein interaction databases to assess the accuracy of extracted events.
• Interactions are scored based on confidence levels derived from database comparisons.

Main Results:

• FLUTE reliably determines the confidence of biomolecular interactions extracted by machine readers.
• The tool significantly speeds up the interaction filtering process, achieving a three-orders-of-magnitude improvement.
• Accurate filtering enhances the reliability of computational model assembly.

Conclusions:

• FLUTE provides a robust solution for filtering inaccurate biomolecular interactions from large-scale literature mining.
• The tool facilitates rapid and accurate assembly of computational models by ensuring data quality.
• FLUTE represents a significant advancement in processing and utilizing biomedical literature for systems biology.