Skip to main content
Contact Info
Hyun Kil Shin
Toxicoinformatics, Department of predictive toxicology, Korea Institute of Toxicology, Daejeon, South Korea

Hyun Kil Shin is currently a senior researcher in Korea Institute of Toxicology and working on the projects to develop machine learning (ML) model for safe compound design. His research  focus is on application of advanced ML algorithms on molecular structure datasets. Particularly  he is interested in development of novel molecular representation applicable to deep learning  model. 

During his research career, Dr. Shin has authored 15 peer-reviewed research articles and book  chapters and served as a reviewer in seven SCI journals (e.g., journal of medicinal chemistry).  He joined university of science and technology (UST) as an associate professor since 2023 at  the human and environmental toxicology department.

OpenTox Virtual Conference 2023

ToxSTAR: prediction of drug-induced cholestasis, cirrhosis, hepatitis, and steatosis

Prediction of Drug-Induced Liver Injury (DILI) is still a challenging task. Each indication,  including cholestasis, cirrhosis, hepatitis, and steatosis, have distinct underlying mechanisms,  making DILI prediction complex. Moreover, drug metabolites, produced through drug  metabolism in the hepatocyte, can exhibit varying toxicity profiles compared to the parent drug.  To address these challenges, the development of New Approach Methods (NAMs) utilizing in  vitro assays and in silico modeling has emerged as a promising solution to enhance DILI  prediction. These NAMs leverage advanced technologies to bridge the gaps in traditional  toxicity testing and provide more precise evaluations, supporting early identification and  mitigation of potential DILI risks. 

ToxSTAR offers a user-friendly web interface that enables users to input drug molecules or their metabolites for prediction of drug-induced cholestasis, cirrhosis, hepatitis, and steatosis. It  provides (Q)SAR model prediction values based on the molecular structures and hepatotoxicity  analysis based on structural similarity between query molecules and drug molecules in the database, which contains in vitro assay and in silico prediction data. At the moment, ToxSTAR  is available in desktop version; however, it will be soon released in mobile application (url: