A unified computational framework for quantitative design and optimization of transcriptional regulation across bacterial species

Precise modeling of transcriptional regulation is essential for the rational design of genetic circuits in synthetic biology. Current computational approaches for predicting transcriptional activity (ITX) typically lack mechanistic clarity, composability, and scalability, and require extensive training data. Here, we present a modular thermodynamic modeling framework that explicitly parameterizes molecular interactions among promoters, RNA polymerase (RNAP) and transcription factors (TFs). Implemented as the computational platform, T-Pro, this approach provides robust interpretability, scalability, and predictive power. Experimental validation across three distinct bacteria-Escherichia coli, Bacillus subtilis, and Corynebacterium glutamicum-demonstrates substantial improvements (up to 20-fold) in a composite transcriptional performance metric (Fmax*FC), achieved within only three Design-Build-Test-Learn cycles and fewer than five genetic constructs in total. Furthermore, we validate the framework by engineering multispecies bacterial communication circuit, highlighting its broad utility and generalizability. The principles and tools developed here thus enable efficient, rational optimization of transcriptional regulation across diverse prokaryotic hosts.