iProLINK: an integrated protein resource for literature mining

Area of Science:

• Bioinformatics
• Computational Biology
• Molecular Biology

Background:

• Exponential growth in molecular sequence data and scientific literature necessitates advanced biological literature mining.
• Existing text mining methodologies lack adequate curated data for training and benchmarking.
• The Protein Information Resource (PIR) curates the UniProt protein sequence database.

Purpose of the Study:

• To develop iProLINK, an integrated Protein Literature INformation and Knowledge resource.
• To provide curated data sources for text mining research in protein annotation.
• To support bibliography mapping, annotation extraction, protein named entity recognition, and protein ontology development.

Main Methods:

• Developed iProLINK, a resource for protein literature mining.
• Created mapped citations linking PubMed IDs to protein entries and feature lines.
• Compiled annotation-tagged literature corpora, including abstracts and full-text articles with experimentally validated post-translational modifications (PTMs).
• Assembled a protein name dictionary, word token dictionaries, and protein name-tagged literature corpora with tagging guidelines.
• Established a protein ontology based on PIRSF protein family names.

Main Results:

• iProLINK provides curated data sources for various text mining applications.
• The resource includes mapped citations and tagged literature corpora for bibliography mapping and annotation extraction.
• iProLINK offers resources for entity recognition and ontology development, including a protein name dictionary and a protein ontology.

Conclusions:

• iProLINK addresses the need for curated data in biological literature mining.
• The resource facilitates research in protein annotation, named entity recognition, and ontology development.
• iProLINK is freely accessible, promoting advancements in bioinformatics and computational biology.