TCRE

TCRE extracts relationships between cell types, cytokines, and transcription factors from biomedical literature to characterize molecular signaling networks underlying T cell function and differentiation.

Key Features:

• Large-scale relation extraction: Extracted 283,000 relations involving cell types, cytokines, and transcription factors from a corpus of 64,000 documents (53,000 PubMed Central full-text articles and 11,000 PubMed abstracts).
• Biological focus: Targets molecular interactions relevant to T cell function, differentiation, and plasticity.
• NLP methodologies: Employs weak supervision and transfer learning for relation extraction from text.
• Result validation: Extracted relations are validated against known biological interactions.

Scientific Applications:

• Mechanistic insights: Enables identification of cytokine and transcription factor interactions that inform mechanisms of T cell plasticity and function.
• Hypothesis generation: Provides extensive relational data to generate testable hypotheses about T cell behavior and potential therapeutic targets.
• Data integration: Facilitates integration of literature-derived relations with experimental findings to support broader analyses of T cell biology.

Methodology:

Corpus compilation of a large-scale biomedical literature collection; relation extraction using NLP techniques including weak supervision and transfer learning; validation of extracted relations against known biological interactions.