AOP-helpFinder

AOP-helpFinder identifies and extracts associations between environmental stressors and biological events (molecular initiating events, key events, and adverse outcomes) to support adverse outcome pathway (AOP) development and chemical risk assessment.

Key Features:

• Text mining (natural language processing): Extracts relevant information from scientific literature using natural language processing techniques.
• Graph theory (network construction): Constructs networks that represent associations between identified terms to enable analytical exploration of potential pathways.
• PubMed screening: Screens over 30 million PubMed abstracts to identify co-mentioned terms such as stressors and molecular initiating events (MIEs).
• Artificial intelligence / machine learning: Uses machine learning algorithms to automate and enhance identification of associations between stressors and key events (KEs).
• AOP and adverse outcome network construction: Supports building adverse outcome pathways (AOPs) and adverse outcome networks (AONs) to inform chemical risk assessment and identification of potential effects such as endocrine disruption.

Scientific Applications:

• Toxicology and environmental health sciences: Enables literature-based identification of stressor-event relationships for mechanism-based assessment and AOP development.
• Chemical risk assessment: Provides evidence to inform chemical hazard evaluation and the identification of potential adverse outcomes associated with chemicals, including endocrine disruptors.
• Alternative testing strategies and predictive modeling: Supplies data-driven associations to support alternatives to traditional toxicological testing and the development of predictive models for adverse outcomes.

Methodology:

Text mining using natural language processing; screening of >30 million PubMed abstracts for co-mentioned terms; construction of association networks via graph theory; and application of machine learning algorithms to identify and prioritize associations.