University of Seoul Develops AI Toxicity Prediction Model as Alternative to Animal Testing

A New Paradigm in Toxicity Assessment
Explainable AI Models Pave the Way for Animal Testing Alternatives
University of Seoul Team Publishes Breakthrough Research in Computational Toxicology

The University of Seoul announced on August 27 that Professor Jinhee Choi and her research team from the Department of Environmental Engineering have developed an artificial intelligence (AI)-based toxicity prediction model and, based on this achievement, have laid the foundation for the development of explainable AI-based toxicity prediction models.

Image of Professor Jinhee Choi's paper. University of Seoul

Two papers detailing the results of this research were published online in the internationally renowned journal "Computational Toxicology" on July 17 and 25.

Globally, the 3R principles of "Replacement, Reduction, and Refinement" for animal testing in the bio field are being strengthened, and AI-based toxicity prediction models are becoming a core alternative to animal testing methods.

The explainable AI-based toxicity prediction model being developed by Professor Choi's research team is designed for use in regulatory decision-making and is considered a key technology that could revolutionize the paradigm of chemical toxicity assessment in the future.

The research team built the AI model using in vitro cell and molecular-level data from the ToxCast database operated by the U.S. Environmental Protection Agency (EPA), as well as in vivo toxicity data from the global chemical information search platform eChemPortal, provided by the Organisation for Economic Co-operation and Development (OECD).

Furthermore, by applying the Adverse Outcome Pathway (AOP) framework to integrate these models, the team is completing the development of an explainable toxicity prediction model.

Professor Choi emphasized, "With the advancement of cutting-edge data science technologies, the paradigm in the field of toxicology is shifting, and explainable AI-based toxicity prediction models will become an important technology for public safety and can be applied across various fields in the future."

Meanwhile, this research was supported by the Ministry of Environment's "Core Technology Development Project for Prevention and Management of Environment-Related Diseases" and the "Specialized Graduate School for Chemical Safety Management."

This content was produced with the assistance of AI translation services.