Design Principles of Life, Disease, and Aging

# Where the Alon Lab is going now: tissues, aging, and predictive medicine *Evidence guide: Current research results under modeling assumptions. Predictions and population heritability need appropriately limited interpretation.* First, healthy years. Many longevity interventions in animals stretch the whole survival curve. A 2025 study by Yang and colleagues used the saturating removal model to argue that such stretching also stretches the sick period, while interventions that steepen the survival curve could compress the fraction of life spent sick. The support comes from mice, worms and flies. Second, genetics. How much of the variation in human lifespan is heritable? Twin studies have estimated twenty to twenty-five percent, and some large family studies even less. In a 2026 Science paper, Shenhar, Alon and colleagues argued that these estimates were biased downward by extrinsic mortality: deaths from infections, accidents and other outside causes, which were common for people born around 1900, and which mask the genetics of biological aging. They used mortality models, including the saturating removal model, to correct for extrinsic deaths in Danish and Swedish twin cohorts, including twins raised apart. The estimated heritability of intrinsic lifespan rose to about fifty percent. What does that mean? Heritability is a population statistic: the share of variation in lifespan, within a particular population and environment, associated with genetic differences. It does not mean genes determine half of how long any one person will live. The rest of the variation reflects environment, chance and other factors, and the authors stress that heritability can change with the population and the era. In a 2026 Nature paper, Somer, Mannor and Alon introduced one-shot tissue dynamics reconstruction, or OSDR. Spatial proteomics maps each cell's type and position, and a marker called Ki-67 shows which cells are dividing. Different neighborhoods in the same tissue have different compositions. OSDR learns how the composition of a cell's neighborhood predicts its chance of dividing. From that, it builds a dynamical model, and a phase portrait of how the cell populations would change over time. Applied to more than seven hundred breast cancer biopsies, OSDR reconstructed fibroblast and macrophage dynamics with hot and cold fibrosis states, in agreement with co-culture experiments. It also suggested an excitable circuit in which T cells can trigger a pulse of T and B cells, like an immune flare. In a clinical trial in triple-negative breast cancer, with biopsies before, early during, and after treatment, models fit to early-treatment biopsies predicted a collapse of the tumor-cell population in responders, but not in non-responders. A biopsy does not literally record the future. OSDR infers a model under assumptions, such as death rates that do not depend on the neighborhood, and dynamics that stay roughly constant. Its predictions need validation. Put the whole series together: complex data, a simple circuit, mathematical dynamics, an experimental test, a design principle, its fragility, and a prediction about disease or aging. Sources: [yang2025](https://doi.org/10.1038/s41467-025-57807-5), [shenhar2026](https://doi.org/10.1126/science.adz1187), [aging-notes](https://www.weizmann.ac.il/mcb/alon/courses/system-biology-aging-and-longevity-2026), [somer2026](https://doi.org/10.1038/s41586-025-09876-1).