An image segmentation, lineage and analysis tool for bacteria: Applications in cell proliferation and cytoplasmic dynamics.

Abstract

Because of the stochastic nature and significant cell-to-cell variation of many biological processes, it is essential to analyze a significant number of cells to understand such processes at the single cell level. Such quantitative biological questions require fast yet reliable automated computational algorithms to analyze images containing thousands of bacterial cells. We present an automated image processing package well-suited to quantitative analysis of high-throughput live-cell fluorescence microscopy of bacterial cells. We describe several capabilities of our software, including identifying and linking cells from frame-to-frame, and characterizing the cell morphology and fluorescence. In particular, we demonstrate its strong performance in reliably segmenting micro-colonies with many cells, facilitating the analysis of cell-cycle dynamics in bacteria as well as cell-contact mediated phenomena. We apply the package in two separate quantitative biological questions. We investigate cell proliferation with single cell resolution during lag phase, the transition of non-growing cells to a growing state. By analyzing the generation times of growing and non-growing progenitors, we suggest a qualitative model for lag phase. As a second application, we characterize the anomalous dynamics in the cytoplasm by quantitative analysis of thousands of complete cell-cycle fluorescence trajectories of molecular complexes. Our results support two modes of nucleoid action: volume exclusion in organizing the cell and a mode of rapid motion. We also report emergent self-similar structure in the dynamics of the cytoplasm and propose a general mechanism by which scale independence emerges in the strong-disorder limit of biological systems.