Bio:　Dr. Chao Ji is a staff scientist at the Agency for Toxic Substances and Disease Registry (ATSDR), Centers for Disease Control and Prevention (CDC), where she applies and develops in silico tools to analyze toxicity data, address toxicity data gaps, and provide guidance on environmental health and chemical risk assessments to support public health policy decisions. Her work includes analyzing toxicity data using benchmark dose (BMD) modeling and providing clear, actionable interpretations and conclusions for the ATSDR Minimal Risk Level Workgroup. She earned her PhD from the University of Miami, followed by postdoctoral training in public health at Indiana University Bloomington, where she contributed to the development of the genomic module of BBMD.

Abstract: The number of chemicals introduced into the environment has substantially exceeded the capacity of traditional risk assessment. Consequently, the next generation of risk assessment involves the utilization of high-throughput transcriptomics to evaluate chemical safety. Existing studies have revealed that BMD estimates from short-term in vivo transcriptomics studies can approximate those from long-term guideline toxicity assessments.

This talk aims to provide the audience with an understanding of BMD estimation from toxicogenomics data. We will focus on the standard toxicogenomics BMD analysis procedure following the National Toxicology Program's approach to genomic dose-response modeling. Specifically, responsive genes are identified using statistical methods and are fitted with parametric dose-response models to determine biological potency estimates. With the estimated BMDs, genomic points of departure, with biological and mechanistic interpretations, are derived by grouping genes into pathways. The existing publicly available toxicogenomics data analysis software, including BMDExpress and Bayesian BMD (BBMD), will be introduced and compared. BMDExpress utilizes maximum likelihood estimation to select the “best” model for biological potency, while BBMD uses Bayesian model averaging to account for uncertainty. A demonstration of BBMD software will be conducted through case studies utilizing publicly available datasets from the Open TG-Gates database. This talk will provide participants with the skills to conduct genomic BMD analyses, a vital tool for advancing risk assessment and safeguarding public health.