Date Approved
9-8-2025
Embargo Period
9-8-2027
Document Type
Dissertation
Degree Name
Ph.D. Pharmaceutical Chemistry
Department
Chemistry and Biochemistry
College
College of Science & Mathematics
Advisor
Hao Zhu, Ph.D
Committee Member 1
Zhiwei Liu, Ph.D.
Committee Member 2
Zhihong Wang, Ph.D.
Committee Member 3
Lei Yu, Ph.D.
Committee Member 4
Jinglin Fu, Ph.D.
Keywords
Cheminformatics;Machine learning;Public database;QSAR;Toxicity
Disciplines
Chemistry | Physical Sciences and Mathematics
Abstract
The pharmacological activity and toxicity of chemical entities are critical factors influencing their success in drug discovery and development. Traditionally, in vivo animal studies have been employed to evaluate their efficacy and safety profiles. However, these studies are often costly and time-consuming. Artificial intelligence (AI)-powered assessments have emerged as promising alternatives for evaluating the bioactivity and toxicity of both new drugs and drug candidates. This dissertation presents novel AI-driven modeling approaches that leverage machine learning (ML) to predict the bioactivity and toxicity of small molecules and nanomaterials (NMs). First, we constructed a virtual graphene library using a nanostructure annotation technique and computed novel geometrical descriptors from the annotated nanostructures for ML modeling. The resulting models showed good performance in predicting graphene toxicities. Next, we developed an online nanoinformatics platform that provides annotated nanostructure databases and modeling toolkits for bioactivity and toxicity prediction. This platform serves as a data-driven computational resource, facilitating the rational design of NMs. Lastly, we proposed a new ML framework integrating drug pharmacokinetic (PK) data to predict placental drug permeability and assess fetal exposure risks. Key transporters identified by the model reveal critical routes of drug transfer across the placental barrier.
Recommended Citation
Wang, Tong, "ARTIFICIAL INTELLIGENCE-POWERED DRUG DISCOVERY: PREDICTIVE MODELS FOR THE BIOACTIVITY AND TOXICITY OF SMALL MOLECULES AND NANOMATERIALS" (2025). Theses and Dissertations. 3449.
https://rdw.rowan.edu/etd/3449
