Mutually inhibiting teams of nodes: A predictive framework for structure–dynamics relationships in gene regulatory networks

Abstract

Phenotypic plasticity—the reversible switching of cell-states—is a central tenet of development, regeneration, and cancer progression. These transitions are governed by gene regulatory networks (GRNs), whose topological features strongly influence their dynamics. While toggle switches (mutually inhibitory feedback loops between two transcription factors) are a common motif observed for binary cell-fate decisions, GRNs across diverse contexts often exhibit a more general structure: two mutually inhibiting teams of nodes. Here, we investigate the teams of nodes as a potential topological design principle of GRNs. We first analyze GRNs from the Cell Collective database and introduce a metric, impurity, which quantifies the fraction of edges inconsistent with an idealized two-team architecture. Impurity correlates strongly with statistical properties of GRN phenotypic landscapes, highlighting its predictive value. To further probe this relationship, we simulate artificial two-team networks (TTNs) using both continuous (RACIPE) and discrete (Boolean) formalisms across varying impurity, density, and network size values. TTNs exhibit toggle-switch-like robustness under perturbations and enable accurate prediction of dynamical features such as inter-team correlations and steady-state entropy. Together, our findings establish the teams paradigm as a unifying principle linking GRN topology to dynamics, with broad implications for inferring coarse-grained network properties from high-throughput sequencing data.