ModEx

ModEx annotates transcription factor–gene regulatory interactions with mode-of-regulation metadata (activation or repression) by extracting evidence from PubMed to augment ChIP-seq interaction maps.

Key Features:

• Text Mining Capabilities: Leverages text mining of PubMed articles to extract literature evidence linking transcription factors to target genes.
• Mode of Regulation Annotation: Classifies TF-gene interactions as activators or repressors based on extracted textual evidence.
• Performance Evaluation: Evaluated against curated databases, achieving an F-score of 0.77 on TRRUST-curated interactions and an F-score of 0.96 on the intersection of TRUSST and ChIP-network datasets.

Scientific Applications:

• Understanding Regulatory Pathways: Annotates TF-gene interactions with mode of regulation to aid deciphering regulatory networks underlying development and diseases such as cancer.
• Machine Learning Models: Provides annotated interaction data for training machine learning models for biomarker discovery, predicting therapeutic responses, and precision medicine.
• Global Gene Regulatory Mechanisms: Integrates context-specific regulatory information into existing interaction datasets to provide a more comprehensive view of gene regulation.

Methodology:

ModEx applies text-mining techniques to PubMed literature to extract evidence of TF-gene regulatory interactions and annotate interactions derived from ChIP-seq experiments.