Fluctuation Test with Phenotypic Switching: A Unified Stochastic Approximation Framework

This paper examines a structurally symmetric fluctuation test experiment in which cell populations grow from a single cell to a set size before undergoing treatment. During growth, cells may acquire tolerance to treatment through probabilistic events, which are passed to progeny. Motivated by recent research on drug tolerance in microbial and cancer cells, the model also allows tolerant cells to revert to a sensitive state, reflecting dynamic phenotypic switching. The master equation governing the probability distribution of tolerant cells is solved via the generating function method and the quasi-powers approximation. Depending on model parameters, the distribution may be approximated by a stable distribution (or its special case, the normal distribution) or through large deviations theory. In the regime of frequent switching, the large deviations approach provides better agreement with numerical solutions, particularly at distribution tails. Conversely, in the regime of infrequent switching, the general stable distributions offer improved accuracy over the Landau distribution, which represents a limiting distribution in case of unidirectional switching.