In phenotypically heterogeneous microbial populations, the decision to adopt one or another phenotype is often stochastically regulated. However, how this stochasticity affects interactions between competing microbes in mixed communities is difficult to assess. One example of such an interaction system is the competition of an Escherichia coli strain C, which performs division of labor between reproducers and self-sacrificing toxin producers, with a toxin-sensitive strain S. The decision between reproduction or toxin production within a single C cell is inherently stochastic. Here, combining experimental and theoretical approaches, we demonstrate that this stochasticity in the initial phase of colony formation is the crucial determinant for the competition outcome. In the initial phase (t < 12h), stochasticity influences the formation of viable C clusters at the colony edge. In the subsequent phase, the effective fitness differences (set primarily by the degree of division of labor in the C strain population) dictate the deterministic population dynamics and consequently competition outcome. In particular, we observe that competitive success of the C strain is only found if (i) a C edge cluster has formed at the end of the initial competition phase and (ii) the beneficial and detrimental effects of toxin production are balanced, which is the case at intermediate toxin producer fractions. Our findings highlight the importance of stochastic processes during the initial phase of colony formation, which might be highly relevant for other microbial community interactions in which the random choice between phenotypes can have long-lasting consequences for community fate.

Cell migration on microlanes represents a suitable and simple platform for the exploration of the molecular mechanisms underlying cell cytoskeleton dynamics. Here, we report on the quasi-periodic movement of cells confined in stripe-shaped microlanes. We observe persistent polarized cell shapes and directed pole-to-pole motion within the microlanes. Cells depolarize at one end of a given microlane, followed by delayed repolarization towards the opposite end. We analyze cell motility via the spatial velocity distribution, the velocity frequency spectrum and the reversal time as a measure for depolarization and spontaneous repolarization of cells at the microlane ends. The frequent encounters of a boundary in the stripe geometry provides a robust framework for quantitative investigations of the cytoskeleton protrusion and repolarization dynamics. In a first advance to rigorously test physical models of cell migration, we find that the statistics of the cell migration is recapitulated by a Cellular Potts model with a minimal description of cytoskeleton dynamics. Using LifeAct-GFP transfected cells and microlanes with differently shaped ends, we show that the local deformation of the leading cell edge in response to the tip geometry can locally either amplify or quench actin polymerization, while leaving the average reversal times unaffected.

We extend generalized additive models for location, scale and shape (GAMLSS) to regression with functional response. This allows us to simultaneously model point-wise mean curves, variances and other distributional parameters of the response in dependence of various scalar and functional covariate effects. In addition, the scope of distributions is extended beyond exponential families. The model is fitted via gradient boosting, which offers inherent model selection and is shown to be suitable for both complex model structures and highly auto-correlated response curves. This enables us to analyse bacterial growth in Escherichia coli in a complex interaction scenario, fruitfully extending usual growth models.