Doctor of Philosophy in Biology
In the
School of Biological Sciences

Jennifer Farrell

Will defend her thesis

“Infection ecology shapes pathogen dynamics and treatment response”

July 1, 2025
10:30 AM
In-person location: Cherry Emerson 204

https://www.microsoft.com/en/microsoft-teams/join-a-meeting
Meeting ID: 123 456 789 012
Access code: 5J1H0Xd2

Thesis Advisor:
Dr. Sam Brown
School of Biological Sciences
Georgia Institute of Technology

Committee Members:
Dr. Steve Diggle
School of Biological Sciences
Georgia Institute of Technology

Dr. Marvin Whiteley
School of Biological Sciences
Georgia Institute of Technology

Dr. James Gurney
Department of Biology
Georgia State University

Dr. Robert Jerris
Children's Healthcare of Atlanta &
Emory University School of Medicine

 
Abstract: 
Polymicrobial infections are shaped by complex, dynamic ecological forces. These include interactions both between and among species, as well as between species and their environment, each influencing the other. In this dissertation, I examine how microbial interactions and environmental context shape pathogen behavior and treatment response, framing infection as an ecological system rather than a single-species phenomenon. Through a combination of theory and experimentation, I show that factors such as growth medium, inoculum size, and the presence of other microbes significantly challenge standard approaches to infection modeling and treatment.

I begin with a review of chronic polymicrobial infections through an eco-evolutionary lens, setting the stage for empirical work examining how microbial dynamics are shaped by ecological context. I then investigate causal relationships between viral and secondary bacterial infections, and ensuing mortality. Using clinical data from the COVID-19 pandemic alongside causal simulation modeling, I show that the widely accepted model of bacterial mediation of mortality is not the only causal model consistent with the data. Drawing on insights from past viral respiratory pandemics, I demonstrate that alternate causal models have substantially different implications for infection management, and identify the types of data needed to disentangle bacterial and viral contributions to morbidity and mortality. 
I then shift to in vitro experimental work exploring how nutritional and microbial context affect pathogen survival and antibiotic response. First, I examine how growth medium and inoculum density combine to modulate antibiotic susceptibility, comparing standard MIC metrics with alternative measures, including a novel cluster-based approach I term KT (k-threshold). I find that susceptibility is highest in Mueller-Hinton medium, the clinical standard, but significantly reduced in standard research laboratory medium (LB) and in synthetic cystic fibrosis medium (SCFM1), which more closely resembles the nutrient conditions of in vivo infections. Importantly, I find that inoculum effects are substantially stronger in LB and SCFM1 compared to the clinical standard MH, raising concerns about ecological realism in standard testing protocols. I then extend the experimental investigation of infection complexity by asking how a commensal competitor affects pathogen growth from rare, and how their interaction both influences and is influenced by antibiotic treatment. I develop theoretical predictions, identify Neisseria subflava as a suppressor of rare Pseudomonas aeruginosa, and show that while either N. subflava or antibiotic exposure alone suppress P. aeruginosa, their combination can abolish or reverse suppression, resulting in competitive release.
Finally, I conclude with a discussion focusing on two central questions that emerged from my work. First, I re-examine definitions of pathogens and infections within the context of polymicrobial systems; second, I address what constitutes a useful and effective experimental infection model. Using the ‘uncanny valley’ concept from robotics, I propose that similar phenomena can apply in experimental infection research, where the addition of more ‘human’ components to an infection model does not necessarily result in experimental models with greater utility.