PIE the search

PIE the search extracts protein-protein interaction (PPI) information from biomedical literature using machine learning-based word and syntactic analyses to prioritize and score articles by PPI confidence for downstream curation and research use.

Key Features:

• Protein-protein interaction extraction: Identifies PPI mentions and related evidence within biomedical literature.
• Machine learning-based text analysis: Applies machine learning techniques for word-level and syntactic analyses of text.
• Article prioritization: Prioritizes PPI-informative articles to focus curation and review efforts.
• PPI confidence scoring: Assigns PPI confidence scores to articles to enable ranked retrieval of likely PPI-containing publications.
• Literature sources: Operates on biomedical literature collections including PubMed abstracts and related texts.
• Competition-winning approach: Implements an algorithmic approach described as competition-winning to improve PPI detection performance.

Scientific Applications:

• Literature curation: Supports curators in identifying and prioritizing articles for manual extraction of PPI data.
• Research support: Provides researchers and biologists with prioritized PPI evidence to inform experimental design and hypothesis generation.
• Database augmentation: Facilitates discovery of up-to-date interaction information not yet represented in manually curated databases.

Methodology:

Uses machine learning techniques for word and syntactic analyses of text data and applies a competition-winning algorithmic approach to rank articles by PPI confidence.