Spatiotemporal Dynamics of Proteus mirabilis Colonies
Project Description:
Proteus mirabilis, a strain of bacteria that naturally occurs
in the gut, is one of the leading causes of antibiotic-resistant
urinary tract infections in catherized patients. When
a P. mirabilis colony begins migrating (or "swarming"), both
the geometry of the individual bacteria and the shape of the overall
colony undergo drastic changes. Individual cells elongate, align, and
move, aggregating into a variety of higher-level structures as the
leading edge of the swarm moves forward. Formal characterization of
these geometric changes is key to assessing experimental results and
understanding the swarming behavior.
The concentric rings in the image above, which shows
a P. mirabilis colony that has grown outwards from a central
inoculation point, result from the cyclical nature of the swarming
behavior: ~3 hours of doubling, followed by ~2 hours of active
movement, then roughly an hour of cell division before the cycle
repeats. Details: cells from strain BB2000 were incubated at 37°C
on a CM55 blood agar plate containing a dye for better
contrast. Photographs were taken by Andrew Fields with a Canon SLR at
an oblique angle. Owner: Gibbs Lab @ UC Berkeley.
During this cyclical process, the leading edge of the swarm develops a
complex, filigreed structure, as shown in these closeups, taken two
hours apart during an eight-hour experimental run:
We are interested in using computational geometry and topological data
analysis to develop and instantiate formal metrics for characterizing
the important structures in these colonies, which include:
- fingers: densely packed, highly aligned protrusions of elongated
cells emerging from the swarm's leading edge
- flaneurs: isolated bacteria that have moved beyond the leading
edge and lost contact with the bulk of the swarm
- islands: connected subsets of bacteria
A key element in this classification problem is the nature of
communication between
P. mirabilis individuals, which is carried out by exchanges of
proteins through cell-to-cell contact. We have developed a number of
different metrics to get at the effects and implications of this,
including:
- simple adjacency: the fraction of the perimeter of a given cell
that is in contact with the perimeter of another.
- adjacency graphs, with vertices for each cell and edges
defined by cell-to-cell contact
- connected-component analysis on the adjacency graphs
- local packing fraction: the number of bacteria whose midpoints
fall within a given kernel size.
Used separately and together, these metrics can effectively bring out
important aspects of the swarm structure. Below are heatmaps that
color-code (left) individual bacteria according to the fraction of
their perimeter that is in contact with any other bacterium (center)
the local packing fraction (right) the connected components.
In the field of microbiology, there is a pressing need for language
and tools to analyze individual interactions within a bacterial swarm,
particularly those that are surface-bound. Our initial results are
promising, as they suggest that we can successfully define key aspects
of cell behaviors in collective migration. These include cell
adjacency, cell-cell networks, geospatial density maps, and
geometrical descriptions of colony edges, all at a remarkably high
resolution. The potential impact of this research is significant, as
it forms the foundation for future studies that could lead to a deeper
understanding of bacterial behavior and pave the way for new
strategies to disrupt intercellular interactions in mutant strains.
People:
- Liz Bradley,
Profesor, Department of Computer Science, University of
Colorado-Boulder
- Morgan Byers, PhD
student, Department of Computer Science, University of
Colorado-Boulder
- Eliotte Garling, PhD graduate, Department of Plant & Microbial
Biology, University of California, Berkeley
- Karine Gibbs, Associate
Professor, Department of Plant & Microbial Biology, University
of California, Berkeley
- Shally Li, PhD student, Department of Plant & Microbial
Biology, University of California, Berkeley
- Jim Meiss,
Professor, Department of Applied Mathematics, University of
Colorado-Boulder
Papers, talks, and other resources:
- A
fun movie of a P. mirabilis swarm.
-
A
github repository with all of our code and instructions on how to
run it. You can also link to that codebase via Zenodo
via this DOI.
- Papers:
- Posters:
- M. Byers, E. E. Garling, E. Bradley, K. A. Gibbs, and
J. D. Meiss, "The Spatiotemporal Dynamics of Proteus mirabilis
Swarm Cycling," poster at
Dynamics Days 2025, Denver CO.
- M. Byers, E. E. Garling, E. Bradley, K. A. Gibbs, and
J. D. Meiss, "A Multi-Scale Spatial Analysis of Proteus
mirabilis Swarm Dynamics," poster at
Dynamics Days 2026, Tempe
AZ. Best-poster award.
- Talks:
- M. Byers, E. E. Garling, E. Bradley, K. A. Gibbs, J. D. Meiss,
"Geometric and Topological Approaches to Proteus mirabilis
Swarm Characterization" in SIAM Conference on Life Science (LS26),
Cleveland, OH, 2026.
- M. Byers, E. E. Garling, E. Bradley, K. A. Gibbs, J. D. Meiss,
"The Spatiotemporal Dynamics of Proteus mirabilis Swarming" in SIAM
Conference on Applications of Dynamical Systems (DS25), Denver, CO,
2025.
- M. Byers, E. E. Garling, E. Bradley, K. A. Gibbs, J. D. Meiss,
"Spatiotemporal Analysis of Proteus mirabilis Swarming” in the
Department of Mathematics BioMath Seminar Series, William & Mary,
Williamsburg, VA, 2025.
Support:
