Size-structured populations with growth fluctuations: Feynman–Kac formula and decoupling
Ethan Levien, Yair Heïn, Farshid Jafarpour · 2025
Read Size-structured populations with growth fluctuations: Feynman–Kac formula and decoupling by Ethan Levien, Yair Heïn, Farshid Jafarpour (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-14680.txt.
Introduction
Ethan Levien, Yair Hein, and Farshid Jafarpour present a 2025 math biology paper on cell groups that grow at uneven rates. In their model, each cell grows based on a changing inner trait, such as gene activity. Earlier models had shown that this trait can sometimes act apart from cell size.
The authors give general rules for when that split holds, in both single cell lines and whole groups. They link the rules to the Feynman-Kac formula, a known tool for such growth problems. When the split holds, size growth can be recast as a simpler, even process, which makes key values easier to compute.
Catalog record
| CREATOR | Ethan Levien, Yair Heïn, Farshid Jafarpour |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 72 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-14680
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Size-structured populations with growth fluctuations: Feynman–Kac formula and decoupling Ethan Levien, Yair Heïn, Farshid Jafarpour Abstract We study a size-structured population model of proliferating cells in which biomass accumulation and binary division occur at rates modulated by fluctuating internal phenotypes. We quantify how fluctuations in internal variables that influence both growth and division shape the distribution of population phenotypes. We derive conditions under which the distributions of size and internal state decouple. Under this decoupling, population‑level expectations are obtained from lineage‑level expectations by an exponential tilting given by the Feynman–Kac formula. We further characterize weaker (ensemble‑specific) versions of decoupling that hold in the lineage or the population ensemble but not both. Finally, we provide a more general interpretation of the tilted expectations in terms of the mass-weighted phenotype distribution. ∗ These authors contributed equally to this work. Contents 1 Introduction In modern biology single-cell data are abundant, owing largely to microfluidic devices such as the mother machine [ 50 ] , which enable precise measurements of gene expression and morphology within single-cell lineages. Traditionally, phenotype measurements (e.g., cell size, gene expression levels) were predominantly conducted in growing populations. Consequently, understanding how single-cell growth and division dynamics shape phenotype distributions in large populations has become essential. Specifically, in growing populations (figure 1 C), selection biases samples toward faster-growing cells, causing systematic differences between lineage-level and population-level distributions. Understanding the relationship between these distributions is thus critical, as it (1) reveals how populations respond to selective pressures and (2) allows inference of selection strength by comparing single-cell with population-level data [ 29 , 25 , 30 ] . There is a vast literature on these questions, which can be traced back to foundational work by Euler [ 3 ] . In the mathematics literature, work on these topics falls under the umbrella of structured population models. Sharpe and Lotka were the first to study branching models with age dependence [ 45 ] . Powell later explicitly connected these models with experimental measurements of microbial population growth [ 40 ] . Many others have explored mathematical questions, such as existence and uniqueness, primarily using tools from semigroup theory [ 13 , 51 , 52 ] . The importance of size regulation [ 47 ] for explaining experimental measurements was only recently appreciated [ 31 ] and has motivated considerable efforts in the biophysics community [ 31 , 37 , 54 , 23 , 39 , 18 ] . These studies focus separately on the dynamics of a fluctuating phenotype or on the mechanisms of cell-size regulation. In this paper, we study a model of size-regulated growth in which the growth rate is determined by internal variables {\boldsymbol{X}}\in\mathbb{R}^{d} (e.g. expression levels) that evolve independently of cell size between divisions, and may be affected at division. We further assume that {\boldsymbol{X}} may affect the division rate, denoted \beta . Within this model, we explore (1) the structure of the stationary distribution of phenotypes in both the lineage and population ensemble, and (2) representations of population statistics in terms of lineage statistics, which are based on the Feynman–Kac formula. Previous work on Feynman–Kac representations has focused on models lacking the size variable [ 35 ] , but a basic fact is that the division rate \beta cannot be independent of size. This is most easily seen in a model where \lambda is fixed (see section 4.1 ). In the absence of any feedback from size to the division rate, the log size will undergo a random walk, and therefore the variance in sizes will diverge [ 1 ] . Meanwhile, the dependence of \lambda and \beta on {\boldsymbol{X}} generates a potentially complex correlation structure between size and the phenotype {\boldsymbol{X}} within a population. This makes the problem of determining the distribution of {\boldsymbol{X}} difficult, because it requires marginalization over the size distribution. Interestingly, recent work [ 20 ] has shown that in certain cases the size and growth rate are asymptotically decoupled, which means that they are independent variables in a snapshot of the population or a collection of lineages. This phenomenon was also observed in [ 48 ] , in which the authors study the fitness effects of gene expression noise. When the growth rate and cell size decouple, we can study the growth rate distribution in a simpler model without a size variable. This has many practical advantages, such as making inference procedures robust to the model of cell size and making simulations more straightforward. To show that size and growth decouple, in [ 20 ] , the authors derive a lineage representation of population statistics. This has the form of a tilted expectation, which, as we will discuss, is closely related to the Feynman–Kac formula. In the model studied in [ 20 ] , division is prescribed by a random threshold that may depend on the previous cell sizes. Our motivation was to understand more generally the relationship between the size and phenotype distributions. This is achieved by building upon the adder-sizer model in [ 53 , 2 , 24 ] by combining it with the fluctuating growth rate model studied in [ 20 , 28 , 48 , 21 ] . We specifically sought to answer the following questions: When do the size and phenotype asymptotically decouple in either the lineage or population distributions? When decoupling occurs, it is straightforward to represent the population statistics as a tilted expectation, which does not involve the size phenotype. This motivates another question: What is the interpretation of this tilted expectation when decoupling does not occur? This leads us to a generalized Feynman–Kac relationship, which relates lineage statistics to a mass-weighted phenotype distribution. Figure 1: (A) A diagram of the model. A cell has size ( Y ) and growth ( X ) phenotypes. Division occurs at a rate that depends on both. We have omitted the explicit dependence on the initial size in this figure for simplicity. (B) A simulation of our model in the case when X is an OU process, showing the evolution of X (blue) and size Y (red). (C) A diagram of a growing population showing the distinction between the lineage and population distributions. 1.1 Organization of this paper In section 2 , we introduce a general model for the dynamics of size-regulated single-cell growth, and we provide PDE descriptions of the population and lineage distributions. In section 3 , we present our results about this model without much technical detail, while section 4 is devoted to some examples and numerical simulations. In section 5 , we show how the population distribution can be related to the lineage distribution via the Feynman–Kac formula when the size and growth phenotype decouple. This involves using a time change to “tilt”, or importance sample, expectations of lineage observables to obtain population observables. In section 6 we provide a more general interpretation of the tilted expectation formula derived from the Feynman–Kac representation and connect our model to the discrete Feynman–Kac formula. In the more general case where size and growth phenotypes do not decouple, this can be understood as a size-biased average. Finally, we present the technical details and proof of our results in section 7 . For ease of reference, table 1 summarizes