UC Berkeley

Collective behaviors enable individuals to aggregate, working together to accomplish tasks, and withstand adverse environments in ways that isolated organisms cannot. Collective behaviors have emerged across the tree of life, necessarily mediated by interactions among individuals. Yet, at the microbial level, the contributions of individual cells to the emergence, propagation, and maintenance of collective behaviors remain unclear. Using Proteus mirabilis cooperative migration (swarming) as a model system, this dissertation combines genetics, computational, mathematical, and microscopy-based methodologies to quantitatively investigate the architecture of bacterial swarm collectives. Quantitative microscopy approaches for resolving individual cell morphometrics within dense, quickly migrating populations stymied formal descriptions of microscale contributions to collective migration. In this dissertation, I introduce Swarmetrics, a quantitative image analysis pipeline designed to tackle dynamically migrating, irregularly shaped, and densely packed bacterial populations. Application of this pipeline over a series of eight-hour time courses with two species of P. mirabilis cells revealed previously unrecognized heterogeneity in cell morphology throughout swarm development, but most notably during active cooperative migration. This single-cell resolution challenged traditional models of P. mirabilis swarm morphology (length, particularly) as a synchronized transition into swarming behaviors. A new model is presented that incorporates population-level morphological variance as a more robust indicator of a P. mirabilis’ place in the cell cycle, comparable to established gene expression markers for swarming. Overall, this work demonstrate that some collective behaviors may be more heterogenous and dynamically structured than once believed. Building upon these findings, I joined a collaborative team of computer scientists and mathematicians at CU Boulder to establish and characterize the population architecture of swarms at the mesoscale. Together, our efforts have yielded new tools for dissecting complex microbial communities in a quantitative manner. By comparing cell-cell contact and the clustering of cells of wild type and strains deficient in the ability to transfer the non-lethal molecularly encoded identity signal Ids, our research suggests that identity signaling may serve as a mechanism to organize individuals into collectives. Taken together, the work presented in this dissertation proposes that collective behaviors are not homogenous masses of identical cells that coordinate in perfect unison. But instead, the amalgamation of heterogenous and dynamically structured into coordinated collective groups. The approaches developed within this body of work mark a step forward in linking individual-level behaviors to more broad population-scale behaviors leading to the emergence of collectivity.