FLUTE

FLUTE filters and scores biomolecular interactions extracted from biomedical literature to prioritize high-confidence protein interactions for computational modeling.

Key Features:

• Filtering and Scoring: Filters and scores machine-extracted biomolecular interactions for accuracy.
• Cross-referencing with Public Databases: Cross-references machine-extracted events with public protein interaction databases such as those available through PubMed to assess reliability.
• High-throughput Processing: Processes outputs from state-of-the-art machine reading methods that can produce hundreds of thousands of events.
• Performance Speedup: Achieves a speedup by three orders of magnitude in the filtering process compared to traditional methods.
• Confidence Assessment: Rapidly assesses confidence levels in extracted interactions to prioritize high-confidence data.
• Model Assembly Support: Facilitates assembly of precise computational models by providing high-confidence biomolecular interactions.
• Integration of Machine Learning and Databases: Combines machine learning efficiency with database-based reliability to streamline filtering.

Scientific Applications:

• Knowledge Retrieval: Enhances retrieval of biomolecular interaction information from biomedical literature.
• Systems Biology: Supports assembly of precise computational models in systems biology using prioritized protein interactions.
• Hypothesis Generation and Testing: Accelerates hypothesis generation and testing by supplying curated, high-confidence interaction sets.
• Understanding Complex Systems: Contributes to deeper understanding of biomolecular networks and complex biological systems.

Methodology:

Ingests events from machine reading methods, cross-references those events against public protein interaction databases (e.g., PubMed), filters and scores interactions for accuracy, and computes confidence levels for retained interactions.