System-level reorganization of gene networks underlies the nurse-to-forager transition in honeybees

Division of labor in social insects is enabled by coordinated changes in physiology, neural signaling, and gene regulation, yet it remains unclear to what extent behavioral maturation reflects incremental changes in individual genes and/or a reorganization of transcriptomic regulation at the system level. Here, we analyze the nurse-to-forager transition in worker honeybees ( Apis mellifera ) from a network perspective, treating the transcriptome as a gene–gene correlation system whose global organization reflects collective regulatory states. Using fat body gene expression data across five developmental time points that span an interval of behavioral transition, we construct signed gene–gene correlation networks and quantify their macro- and mesostate properties. We then descend to the level of the gene and construct microstate gene lists of interacting genes. Behavioral maturation of worker honeybees is accompanied by coordinated, nonmonotonic changes in global network structure rather than simple global shifts in interaction strength across the network. Early stages exhibit broadly positive, globally synchronized correlations, whereas later stages show increased modular segregation and the emergence of antagonistic gene–gene relationships. Despite this differentiation, networks remain fully connected with high global efficiency, consistent with a flexible small-world–like organization. Positive correlations primarily support global integration, while negative correlations act as localized, boundary-enforcing constraints. Negative correlations form a structured pattern of antagonism that is sparse and module-specific, occurring primarily between functional gene modules rather than within them. Network layouts and module analyses reveal two polarized, attractor-like regulatory states, a structure recapitulated in a vitellogenin-centered subnetwork and consistent with prior experimental perturbation studies.