主讲人简介：
Dr. Longzhu Shen obtained his Ph.D. degree at Carnegie Mellon University. During the Ph.D. period, his research was centered on studying the reaction mechanisms of a human-made biomimetic catalyst for green applications. His quality research has been highly recognized with a four-year continuous award of the prestigious Mellon presidential fellowship. He also served as the chair of the environmental group of the ACS (American Chemical Society) Pittsburgh local chapter. During his term, he lead his team to win the ChemLuminary award for outstanding sustainability activities. After graduation, he carried on his passion toward green chemistry by joining the Center for Green Chemistry and Green Engineering at Yale. His current active role is leading the research consortium forged by four universities to design safer chemicals with reduced likelihood to incur toxicity.
报告主要内容概述：
Toxicity is a concern with many chemicals currently in commerce, and with new chemicals that are introduced each year. The standard approach to testing chemicals is to run studies in laboratory animals (e.g. rats, mice, dogs), but because of the expense of these studies and concerns for animal welfare, few chemicals besides pharmaceuticals and pesticides are fully tested. Over the last decade there have been significant developments in the field of computational toxicology which combines in vitro tests and computational models. The ultimate goal of this field is to test all chemicals in a rapid, cost effective manner with minimal use of animals. High throughput screening (HTS) methods have emerged as an efficient technology to examine how chemicals disrupt biological pathways and lead to adverse health outcomes. Large data sets (thousands of chemicals, hundreds of measurements per chemical) are ideal for developing machine learning predictive models. One of the challenges to the success of modeling is the construction of a descriptor space. In this talk, I'd like to present a modeling strategy built upon the physical principles. It inherently links the molecular descriptors with the possible molecular initiating events that lead to the toxicity endpoints of interest. The obvious advantage of this approach is the model interpretability and resilience to data noise. I'll use two toxicity endpoints, cytotoxicity and NRF2-ARE pathway perturbations, to illustrate the construction and application of the principle-based toxicity models.
