Inflammatory Bowel Disease Interactome: Abstracts and Component Matrices from 1990-2020

25,971 curated scientific abstracts from PubMed, published between 1990 and 2020, form the core of this corpus. María Emilia Fernández and colleagues used text mining and network analysis to track the occurrence of 1,218 biological components (cells, molecules, genes, processes) across these abstracts. The dataset includes raw and curated abstract tables, target matrices of component synonyms, and a summary matrix of component occurrences.

Use Cases

• Track the evolution of IBD research focus over time based on the 'PublicationYear' metadata and component occurrence data.
• Perform network analysis of biological relationships based on the co-occurrence of 1,218 components across the abstract corpus.
• Validate text-mining pipelines for biomedical literature using the curated set of component synonyms from established reviews.
• Study knowledge construction in a specific disease area based on the longitudinal analysis of scientific abstracts.

Strengths

• Corpus of 25,971 abstracts provides substantial text data for analysis.
• Includes 1,218 distinct biological components mapped with semantic synonyms from authoritative sources like GeneCards and UniProt.
• Explicit temporal coverage from 1990 to 2020 allows for longitudinal trend analysis.
• Data is curated to remove duplicates and abstracts focusing solely on animal models.

Limitations

• Row and column counts for individual tables are unspecified, limiting precise assessment of data scale.
• Column-level documentation is absent; field semantics for metadata must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

PubMed via the Zotero bibliographic manager, with component lists derived from recent review papers.

Collection Method

Text mining of scientific abstracts and network analysis, as described in the associated research paper.

Time Range

1990-2020