Uncovering morphogenetic control strategies with compact recurrent gene-circuit models
July 20
Understanding how multicellular tissues revise their architecture after perturbation is a central goal of developmental biology and regenerative medicine. Normative frameworks such as reaction–diffusion, positional-information gradients and simple feedback laws have yielded valuable insights into pattern formation. However, the minimalism of these formalisms often prevents them from capturing the nuanced, history-dependent trajectories observed in real regenerating systems, encouraging iterative, researcher-dependent tweaks.
Here we introduce a modelling strategy that employs recurrent gene-circuit surrogates to infer the control algorithms driving adaptive morphogenesis. We demonstrate that circuits containing only one to four dynamic regulators consistently outperform classical morphogenetic models—and equal the predictive power of far larger networks—in accounting for tissue remodelling choices across six benchmark regeneration paradigms. Crucially, the learned circuits can be analysed through standard dynamical-systems tools, permitting a unified comparison with existing theories and exposing concrete regulatory motifs that orchestrate surface closure, polarity realignment and proportion regulation.
Our framework also furnishes empirical estimates of the intrinsic dimensionality of pattern control and clarifies the strategies discovered by automated in-silico evolution platforms that optimise regenerative performance. Collectively, we provide a systematic route to uncover interpretable morphogenetic strategies, illuminating the molecular-to-tissue logic of repair and laying groundwork for interventions in both normal development and congenital malformation contexts.