Researchers show that how much different cell lineages share the same environmental ups and downs can change whether it is better for a mother cell to split a conserved protective resource evenly or unevenly between her two daughters. Using a simple model of cell division, the study finds that weak sharing of environmental changes favors equal partitioning. But when a large fraction of environmental “innovations” are shared across lineages, selection can favor a markedly asymmetric split. In the model the favored asymmetric branch appears near a daughter share of about α ≃ 0.2 rather than drifting away continuously from equal division.

The model follows a conserved reserve that a mother cell divides between two daughters. The partition is set by a parameter α (α = 1/2 is equal sharing; α < 1/2 concentrates the reserve in one daughter and leaves the other poor). After division, each lineage experiences a sequence of environmental states drawn from two possibilities: a favorable state (R) that recharges the reserve and an adverse state (P) that depletes it. Reserve turnover (a leakage term) and costs for keeping reserve are included, and protection in adverse times rises with reserve according to a smooth (logistic) law. The model also adds a phenomenological cost for unequal partitioning so that extreme polarization is not free.

To study long-term success the authors treated each fixed partition rule as producing a random demographic operator that maps the population distribution forward in time under the sequence of environments. They measured the typical long-term growth rate with the top Lyapunov exponent of that operator. By changing the fraction ρ of environmental innovations that are shared across lineages (ρ = 0 means independent environmental draws for different lineages; ρ = 1 means all lineages draw the same global innovation), they showed that increasing shared environmental fluctuations can make asymmetric inheritance advantageous. The paper uses direct evolutionary simulations, operator calculations, and a second-order “dominant-mode” approximation. That approximation separates the asymmetric advantage into a specialization cost in the mean operator and a reduction in sensitivity to collective fluctuations, and it predicts the finite asymmetric branch, while the full operator product sets the exact numerical crossing.