tagtog

tagtog annotates biomedical literature to identify and label biological entities such as gene symbols, gene names, and Gene Ontology terms to support genomic and genetic curation.

Key Features:

• Entity Annotation: Annotates biological entities in full-text articles, including gene symbols, gene names, and Gene Ontology terms.
• Hybrid Annotation Approach: Integrates manual user annotations with automatic machine-learned annotations to improve identification of gene-related entities.
• BioCreative IV Evaluation: Demonstrated extraction of Drosophila melanogaster gene symbols and names from PLOS full-text articles in the BioCreative IV Interactive Annotation Task.
• Scalability: Performance improves with larger annotated corpora, indicating robustness to increasing dataset size.
• Curation Efficiency: Assisted curation reduces the time required compared with traditional manual methods.

Scientific Applications:

• Database curation (FlyBase): Supports continuous updates and accurate data extraction for genomic databases such as FlyBase.
• Drosophila gene mention extraction: Enables identification and extraction of Drosophila melanogaster gene symbols and names from scientific literature.
• Literature-based genomic annotation: Automates parts of the curation workflow to help maintain comprehensive genomic and genetic datasets for research.

Methodology:

Combines manual annotations and automatic machine-learned annotations; the machine learning component is trained on annotated corpora and iteratively improves entity recognition as the dataset expands.