Graduation Year
2026
Document Type
Master's Thesis
Degree
Master of Science
Program
Biological Science
Partner Organization
Buck Institute for Research on Aging
Program Director
Patti Culross, MD, MPH
First Reader
Akos Gerencser
Second Reader
Chad Lerner
Abstract
Traditional methods of hypothesis-based research for pharmacological development are limited in terms of their discovery throughput and ability to develop therapies in the near term. To accelerate the discovery of compounds that specifically promote healthy aging, our objective is two-fold: (1) to develop automated tools for the high-throughput quantification of donor- and compound-specific differences in the cellular morphology of human donor muscle derived myoblasts specifically, and (2) to screen a small-molecule compound library for agents that cause phenotypically aged cells to revert to established youthful phenotypes (referred to as “hits”).
To accomplish this, we developed an automated live-cell painting assay designed to incorporate a specific combination of fluorescent stains to capture morphological features associated with youthful and aged cellular states. Automated protocols were established for assay execution, and a Python-based tool, PickliPy, was created to generate validated instructions for acoustic liquid handling of fluorophores and compound libraries. In parallel, we developed an image-analysis pipeline using a convolutional neural network (CNN) to quantify cellular phenotypes and identify candidate compounds.
Using this platform, we resolved differences in cellular morphology across different human donors and compound treatments. We were also able to detect donor-specific variation in compound sensitivity. Together, these results validate an automated, image-based screening platform that can be used to identify biologically relevant compounds in an aging-context.