Deciphering the directed iron-stress regulatory network of Synechocystis sp. PCC 6803 reveals novel master regulators.

Iron deficiency elicits extensive transcriptional reprogramming in cyanobacteria; however, the directed regulatory hierarchy that coordinates this response remains insufficiently resolved. In this study, we integrated 34 iron-stress transcriptomic samples into a customized deep graph transformer (DGT) framework to infer a directed, context-specific gene regulatory network (GRN) for Synechocystis sp. PCC 6803 under iron limitation. This analysis identified 82 differentially expressed genes, with isiB and the antisense RNA IsrR exhibiting the strongest transcriptional shifts. The resulting GRN displayed a robust scale-free architecture (R2 = 0.95) and clear modular organization, revealing a hierarchical structure governed by 13 master regulators. Functional annotation demonstrated that these regulatory hubs-largely membrane-associated sensors-coordinate pathways involved in iron uptake (slr1319, slr1392), nitrogen assimilation (slr1756), and cell-envelope remodeling (slr0493). Benchmarking against the STRING interaction network confirmed recovery of the conserved core metabolic machinery while uncovering an additional, previously hidden regulatory tier driven by condition-dependent master regulators, most notably sll1242 and sll1862, which remain underrepresented in static interaction resources. Moreover, the DGT model highlighted two overlooked candidates-slr0493 (rfbM-2) and slr1881 (livF)-as putative upstream regulatory drivers. Collectively, these findings yield a high-resolution, directionally informed blueprint of the iron-stress regulon, surpassing the capabilities of traditional static interactomes and offering prioritized targets for metabolic engineering aimed at enhancing stress tolerance in photosynthetic bioproduction systems.