Dragon TF Association Miner: a system for exploring transcription factor associations through text-mining

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Transcription factors (TFs) play crucial roles in gene regulation.
• Understanding TF functional associations is vital for deciphering gene regulatory networks.
• Existing resources may not efficiently link TFs to Gene Ontology (GO) terms and diseases from literature.

Purpose of the Study:

• To develop and present Dragon TF Association Miner (DTFAM), a text-mining system.
• To identify potential functional associations between transcription factors (TFs), Gene Ontology (GO) terms, and diseases using PubMed literature.
• To provide users with comprehensive reports for easier analysis of TF-related biological information.

Main Methods:

• DTFAM employs text-mining techniques on PubMed documents.
• The system was trained and tested on a curated dataset of >3000 PubMed abstracts related to transcription control.
• Performance was evaluated using sensitivity and specificity metrics.

Main Results:

• DTFAM achieved 80% sensitivity and 82% specificity in selecting relevant documents.
• The system generates detailed tabular and graphical reports linking TFs to relevant documents.
• Documents are color-coded for enhanced user inspection and data interpretation.

Conclusions:

• DTFAM effectively extracts and presents TF-GO and TF-disease associations from scientific literature.
• The system complements existing biological resources by revealing subtle connections within biological entities.
• DTFAM simplifies the analysis of large volumes of literature, saving time for researchers studying gene regulatory networks.