DigChem

DigChem extracts evidence sentences describing relationships among genes, chemicals, and diseases from Medline abstracts to support analysis of disease mechanisms and chemical-gene interactions.

Key Features:

• Deep Learning Model: Employs a bidirectional long short-term memory (BiLSTM) network to identify evidence sentences that describe relationships among genes, chemicals, and diseases in Medline abstracts.
• Extensive Database Coverage: Encompasses a database comprising 35,124 genes, 56,382 chemicals, and 5,675 diseases for literature-scale relationship extraction.
• Reliability and Validation: Validates extracted relationships through comparisons with manually curated data and existing databases.

Scientific Applications:

• Drug Discovery: Identifies potential chemical compounds that interact with specific genes implicated in diseases to aid targeted therapy research.
• Genetic Research: Supports exploration of how specific genes may influence disease susceptibility or progression via interactions with chemicals.
• Disease Mechanism Elucidation: Enables mapping of gene-chemical-disease networks to provide insights into molecular pathways involved in disease development and treatment.

Methodology:

Using a BiLSTM-based deep learning model, evidence sentences describing interactions among genes, chemicals, and diseases are extracted from Medline abstracts; these relationships are identified and cataloged; extracted data are validated by comparison with manually curated datasets and existing databases.