Work

Exact ensemble bifurcation analysis reveals hidden bistability and makes signalling dynamics predictable from kinetic parameters

Year: 2026
Type: preprint
Abstract: Whether a signalling motif acts as a graded sensor, a switch or a memory is set by kinetic parameters that are rarely known, yet bifurcation analysis examines one parameter set at a time. We introduce an inverse method that parametrises every steady state of the dual phosphorylation–dephosphorylation cycle by the free kinase-to-phosphatase ratio, giving the complete bifurcation diagram of a parameter set exactly, without initial conditions, in ms. Applied to 40,000 Latin hypercube parameter sets, it yields an ensemble of exact bifurcation diagrams labelled into ten bifurcation classes. Bistability is common (65% of sets) but lies mostly on detached branches that single-start continuation misses (19–27% against 72–88% of asymmetric sets); a closed-form criterion predicts every detached branch. Gradient-boosted classifiers trained on these labels predict the class from kinetic parameters alone (93% accuracy) and rediscover the analytical conditions, identifying measurable rate ratios for designing synthetic switches.
Language: English
Open Access status: green
OpenAlex ID: W7235626386
Cites:
Cited by:
Related to: 10