Dynamics underlying phenotypic variability in cell populations

Cells in a clonal population exhibit variability in shape, size, molecular content and many other phenotypic properties. How this variability reflects the processes of growth, protein production, division and inheritance is an important and still open question. We utilize experimental methods that trace individual bacteria for hundreds of growth and division cycles, to shed light on the dynamic basis of phenotypic variability. Comparing phenotypic distributions over time in single cells and across a cell population, we find a universal distribution shape for cell size and highly-expressed protein content, which collapses on a skewed highly non-Gaussian curve. These findings indicate a buffering of the population level from microscopic processes, and can be explained by a mesoscopic-scale mathematical model of growth, division and homeostasis. In contrast to this universality, time-averaged cell size and protein content remains distinct among lineages over many generations, displaying a form of "ergodicity breaking". This individuality can be traced back to extremely slow dynamics, with typical timescales of more than 100 generations. Correlations among single-cell variables reveal effective homeostasis mechanisms that are also distinct among individual cells, and suggest a global simultaneous homeostasis of multiple phenotype components.