06/17/2026
By Danielle Fretwell

The Francis College of Engineering, Department of Chemical Engineering, invites you to attend a Doctoral Dissertation defense by Zhao Wang titled: "Multi-Scale Modeling for Bioprocess Digital Twins."

Candidate Name: Zhao Wang
Degree: Doctoral
Defense Date: Friday, June 26, 2026
Time: 10 a.m. - noon
Location: Please email the advisor or student for the location.

Committee:

  • Advisor: Seongkyu Yoon, Professor, Chemical Engineering, University of Massachusetts Lowell
  • Dongming Xie, Associate Professor, Chemical Engineering, University of Massachusetts Lowell
  • Murat Inalpolat, Professor, Mechanical and Industrial Engineering, University of Massachusetts Lowell
  • Ketki Behere, Associate Director, Process Validation, Vertex Pharmaceuticals
  • Jacqueline Gonzalez, Associate Director, Upstream Process Development, Alexion Pharmaceuticals

Abstract:
The biopharmaceutical industry relies heavily on mammalian cell culture to produce therapeutic proteins, yet achieving optimal efficiency and stability requires overcoming complex metabolic inefficiencies and formulation limitations. This dissertation presents a comprehensive multiscale framework integrating systems biology, multivariate data analysis, and process engineering to optimize both the production of biotherapeutics.

First, a thorough review of inhibitory byproducts produced during Chinese Hamster Ovary cell culture is presented. This review highlights the metabolic origins of novel inhibitors beyond lactate and ammonia and explores strategies for their mitigation. Building on this foundation, a graphic user interface (GUI) integrating genome scale metabolic modeling and flux balance analysis is developed for the bioprocess digital twin framework. This computational approach enables advanced in-silico design of experiments to rationally modulate nutrient precursors. Consequently, it controls the accumulation of these identified inhibitory metabolites and enhances fed-batch culture performance.

Demonstrating the broader applicability of data-driven modeling to drug product formulation, the research explores the process development and scaleup of lyophilized mRNA lipid nanoparticles. A multivariate data analysis approach utilizing partial least squares regression is employed to systematically screen and optimize lyoprotectant excipients. This approach identifies a robust sucrose and polymer matrix that preserves particle integrity and bioactivity during the freeze-drying process.

Finally, a three-tier multiscale bioprocess digital twin is engineered for continuous mammalian cell culture. By incorporating macroscopic flowsheets, cellular kinetics with parameter estimation via relative root mean square error minimization, and genome scale models, the framework successfully optimizes operational variables. This is demonstrated through a continuous perfusion case study with spent media recycling. The framework defines operational boundaries to minimize the cost of goods while satisfying specific constraints, such as ensuring monoclonal antibody production targets and viable cell densities are met. Together, these integrated methodologies advance the paradigm of digital biomanufacturing and offer predictive, scalable solutions from cell metabolism to final product formulation.