U.S. flag An official website of the United States government
  1. Home
  2. About FDA
  3. FDA Organization
  4. Office of the Commissioner
  5. Office of the Chief Scientist
  6. National Center for Toxicological Research
  7. Science & Research (NCTR)
  8. NCTR Publications
  9. NCTR Research Highlights
  1. NCTR Publications

NCTR Research Highlights

Periodic report on research activities and special events at NCTR

2023 Research Highlights

January 18, 2023

NCTR Artificial Intelligence Initiatives Selected as Winners in the 2022 EMGS Bioinformatics Challenge

Two NCTR scientists—Drs. Ting Li and Xi Chen—have been selected the first- and second-place winners, respectively, for the annual 2022 Environmental Mutagenesis and Genomics Society’s (EMGS) Bioinformatics Challenge! This EMGS Challenge encourages participants to develop novel tools and approaches that use publicly available data to identify signatures of genotoxic hazards or provide insight into their mechanisms of action.

Drs. Li and Chen are working on two of the NCTR-developed AI4TOX four initiatives, SafetAI and AnimalGAN. AI4TOX is an FDA artificial intelligence (AI) program for toxicology under the leadership of Dr. Weida Tong, Director of NCTR’s Division of Bioinformatics and Biostatistics (DBB). The program aims to apply the most advanced AI methods to develop new tools to support FDA regulatory science and strengthen the safety review of FDA-regulated products. 

Dr. Li presented her work on drug-induced liver injury (DILI) “DeepDILI: Deep Learning-Powered Drug-Induced Liver Injury Prediction Using Model-Level Representation” for the EMGS Challenge. Developed by Dr. Li, DeepDILI is an AI drug-safety model within the SafetAI suite that uses a deep learning-powered prediction model designed to use chemical structure information to identify drugs with the potential to cause DILI in humans. The SafetAI initiative aims to develop AI models for toxicological endpoints that are critical to assessing drug safety and may add value to the review of drug candidates prior to human testing.

Dr. Chen presented “AnimalGAN: A Generative AI Alternative to Animal Clinical Pathology Testing” for the EMGS Challenge. Developed by Dr. Chen, AnimalGAN is an AI-based suite that generates specific animal-study datasets for new and untested chemicals by learning from legacy animal-study data. The successful implementation of this AI-based suite is the first example of a generative adversarial network (GAN)-specific application’s use in a virtual animal model. 

Watch the final round of the Bioinformatics Challenge and learn about the other two AI4TOX initiatives, BERTox and PathologAI. For additional information, please contact Dr. Weida Tong.


NCTR Research Highlights Archives

Resources For You

Subscribe to the NCTR Research Highlights E-mail Notifications

Get regular FDA email updates delivered on this topic to your inbox.

Back to Top