INTEGRATING DATA-DRIVEN AND MECHANISM-DRIVEN MODELING IN DRUG DISCOVERY AND HEPATOTOXICITY EVALUATION

Author(s)

Xuelian Jia, Rowan UniversityFollow

Date Approved

9-27-2024

Embargo Period

9-27-2026

Document Type

Dissertation

Degree Name

Doctor of Philosophy (Ph.D.) in 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

Lauren M. Aleksunes, Pharm.D., Ph.D., D.A.B.T.

Keywords

adverse outcome pathway;deep learning;interpretable machine learning;machine learning;mechanistic modeling;QSAR

Subject(s)

Drug development; Hepatotoxicology; Opioids

Disciplines

Chemistry | Medicinal-Pharmaceutical Chemistry | Pharmacy and Pharmaceutical Sciences

Abstract

The pharmacological activities and hepatotoxicity significantly influence the success or failure of new drugs during drug discovery and development. Traditional experimental methods, such as animal models, are costly and time-consuming for chemical testing. There is a great need to develop in vitro and computational modeling to help identify the pharmacological activities and hepatotoxicity potential in the early stage of drug discovery and safety evaluation. In this dissertation, new computational models and associated modeling frameworks were described for predicting the biological profile and hepatotoxicity of chemicals. Firstly, a novel data mining and Quantitative Structure-Activity Relationship (QSAR) workflow was developed to construct a virtual bioprofile for opioids, facilitating the screening of new analgesic opioids. Next, a mechanistic model consisting of structural alerts and in vitro ARE activation was developed to predict chemical hepatotoxicity potentially mediated through the oxidative stress pathway. Lastly, an interpretable deep neural network approach, incorporating in vitro assay results, transcriptome data, and pathway ontology knowledge, was employed to construct a virtual adverse outcome pathway network encompassing various toxicity pathways. This model's interpretation can unveil potential toxicity mechanisms, aiding in identifying chemicals and drugs of concern to human health.

Recommended Citation

Jia, Xuelian, "INTEGRATING DATA-DRIVEN AND MECHANISM-DRIVEN MODELING IN DRUG DISCOVERY AND HEPATOTOXICITY EVALUATION" (2024). Theses and Dissertations. 3300.
https://rdw.rowan.edu/etd/3300