Abstract
Phenotypic heterogeneity is a defining feature of bacterial stress responses, long framed as irreducible noise arising from stochastic molecular events. This review builds on that view, advancing the idea that much of the apparent randomness instead reflects unmeasured determinants-hidden variables-that render cellular behavior predictable once revealed. We survey canonical cases such as antibiotic persistence, oxidative stress resistance, and DNA repair, highlighting how variability once thought to be stochastic can be traced with deterministic factors such as growth, cell-cycle state, microenvironment changes, or molecular inheritance. Cutting-edge tools-from lineage-resolved microfluidics, single-cell RNA-seq, and high-dimensional reporters to machine learning applied to cellular trajectories-are now beginning to expose these hidden layers of causality. By reframing noise as structure awaiting discovery, this perspective sets the stage for a predictive microbiology: one that links the microscopic state of individual cells to emergent population behaviors, with far-reaching implications for understanding pathogenesis and guiding antimicrobial strategies.