PhD Defense by Jennifer Rattray

Primary tabs

In partial fulfillment of the requirements for the degree of


Doctor of Philosophy in Biology

In the

School of Biological Sciences


Jennifer Rattray


Will defend her dissertation:


Collective behavior and morphological complexity in Pseudomonas aeruginosa


Wednesday, August 3rd, 2022

1 PM Eastern Time


In-person: The Dissertation Defense Room, 4th floor of Price Gilbert Library in room 4222
Virtual Link:  https://gatech.zoom.us/j/93202356870

 Thesis Advisor:

Sam P. Brown, Ph.D.

School of Biological Sciences

Georgia Institute of Technology


Committee Members:

Marvin Whiteley, Ph.D.

School of Biological Sciences

Georgia Institute of Technology


Peter Yunker, Ph.D.

School of Physics

Georgia Institute of Technology


Steve Diggle, Ph.D.

School of Biological Sciences

Georgia Institute of Technology


Will Ratcliff, Ph.D.

School of Biological Sciences

Georgia Institute of Technology




In nature, many animals are capable of performing complex behaviors without centralized coordination. A well-studied focus is collective motion in flocks of birds and shoals of fish, both of which are capable of changing collective behavior as a function of individuals responding to their local environment. Similarly, despite their microscopic individual size, groups of bacteria are capable of collectively responding to and restructuring their environment. In this thesis I focus on the gamma proteobacterium Pseudomonas aeruginosa (PA), a well-studied opportunistic pathogen that is known to engage in complex collective behaviors, often controlled by a form of cell-cell communication mediated by diffusible signal molecules called quorum sensing (QS).

First, I query the sensing capacity of QS, quantifying the ability to sense cell density by tracking QS-regulated secreted protease (lasB) expression on the population and single-cell scale. We find that PA can deliver a graded behavioral response (or ‘reaction norm’) to fine-scale variation in population density and show that populations generate graded responses to environmental variation through shifts in the proportion of cells responding and the intensity of responses. Given this ability of PA to quantitatively respond to discrete density environments, we then ask how the molecular machinery of QS shapes the reaction norms to changing density, via signal synthase knockout and complementation experiments. We find that the wildtype reaction norm is robust to the addition of density-independent signal supplements and more broadly, that a positive reaction norm to density is robust to multiple combinations of gene deletion and density-independent signal supplementation.

Switching from QS control of a single gene (lasB), I turn to a complex multigenic and multicellular trait of microcolony growth. Using a collection of diverse environmental and clinical PA isolates, we develop a colony image library of 69 strains in four-fold replication. We then use a combination of image processing techniques to quantify colony morphology and complexity and find that, under common laboratory conditions, morphology and complexity form a robust, repeatable phenotype on the level of individual strains. Based on this replicable visual “fingerprint” per strain, we reasoned that colony image data could be used to classify previously unseen colony images to the strain level. Using a combination of transfer learning and data augmentation we trained a neural network to classify strains, resulting in high-level accuracy (94%). These results indicate that not only do PA strains have characteristic, replicable ‘fingerprints’, but also that these ‘fingerprints’ are learnable and classifiable. These results could provide a basis for predicting other strain-dependent behaviors including virulence or antibiotic resistance. 

Overall, these results highlight that complex and heterogeneous single-cell behaviors can produce robust and consistent patterns on the collective scale of environmental sensing and colony growth.


  • Workflow Status:Published
  • Created By:Tatianna Richardson
  • Created:07/28/2022
  • Modified By:Tatianna Richardson
  • Modified:07/28/2022