Alexandra Dunnum VandeLoo

BioE PhD Defense Presentation 

Date:  Friday, November 20th, 2026 

Time: 12:00-1:00 PM EST 

Location: EBB CHOA Seminar Room (Room 1005) 

Join Online: https://teams.microsoft.com/meet/26248818027208?p=M28QCH1jnKgv6dhl0K 

(Meeting ID: 262 488 180 272 08 | Passcode: To2KL64z) 

  

Committee Members: 

Prof Craig R. Forest, PhD (Advisor) 

   George W. Woodruff School of Mechanical Engineering, Georgia Institute of Technology 

Prof. Levi Wood  

   George R. Woodruff School of Mechanical Engineering, Georgia Institute of Technology   

Prof. Scott Danielsen  

   School of Materials Science and Engineering, Georgia Institute of Technology  

Prof. Julie Champion  

   School of Chemical and Biological Engineering, Georgia Institute of Technology  

Prof. Todd Fernandez  

   School of Biomedical Engineering, University of British Columbia  

 

 

Integration of Experimental Biology and Deep Learning for Protein Engineering and Quantitative Cell Characterization 

Experimental biology relies upon cultured cells for studying mammalian physiology, disease mechanisms, and therapeutic responses. Yet cultured cells are subject to biological variability, and highly manual low-throughput protocols create efficiency bottlenecks and challenges to experimental reproducibility. These challenges motivate the development of approaches like standardized training methods, automation of repetitive tasks, and quantitative analysis of traditionally qualitative outputs. In this work, I identify several such bottlenecks and present artificial intelligence and machine learning-aided workflows to improve the efficiency and reproducibility of single-cell characterization in two areas: single-cell electrophysiology in Aim 1 and single-cell morphology in Aim 2. 

Single-cell electrophysiology techniques such as patch-clamp enable the characterization of diverse cellular processes, from synaptic transmission to ion channel function. The 1991 Nobel Prize recognized discoveries concerning the function of single ion channels, highlighting the significance of these techniques to modern cell biology. In Aim 1, I develop and validate computational and experimental approaches for patch-clamp electrophysiology, apply these approaches to the identification and characterization of ChrimsonR ion channel variants, and investigate the theoretical, experimental, and structural basis of protein language model-guided mutations. The experiments described in Aim 1 depend on cell culture health, which inform experimental protocols to ensure the availability of experiment-ready cells. In Aim 2, I develop and validate computational approaches for quantitative cell segmentation and morphological analysis, and then apply these tools to identify and characterize cell morphological trajectories under changing culture conditions. Together, these aims establish computational and experimental workflows for improving the efficiency, reproducibility, and quantitative characterization of biological systems at the single-cell level.