Natural language processing in toxicology
2024
Utrecht University of Applied Sciences, Utrecht, Netherlands
Toxicologists use Adverse Outcome Pathways (AOPs) to understand how compounds cause harmful effects. AOPs visualize toxicity mechanisms across biological levels. Currently, AOP construction relies on manual literature review, which is time-consuming. Natural Language Processing (NLP) could streamline this process, allowing researchers to focus on critical assessment rather than data gathering. This study used deep learning models to identify relevant entities and causal relationships in scientific literature, focusing on liver cholestasis and steatosis. The NLP pipeline combined Named Entity Recognition and relationship extraction to screen compounds and extract mechanistic information across biological levels. The research demonstrates NLP's potential in supporting toxicological information extraction and provides an open-source model for recognizing toxicological entities and their relationships. Resources are available on GitHub.
The application of natural language processing for the extraction of mechanistic information in toxicology
Marie Corradi
Added on: 12-03-2024
[1] https://www.frontiersin.org/journals/toxicology/articles/10.3389/ftox.2024.1393662/full
