Host-aware Identification of Intrinsic Gene Expression Biopart Parameters using Combinatorial Libraries

Model-based design in synthetic biology is limited because bioparts are typically characterised by relative metrics that vary across genetic and physiological contexts. To address this, we introduce a host-aware framework for quantitatively characterising bioparts in combinatorial libraries of plasmid-based constitutive expression constructs. The approach integrates a digital twin of Escherichia coli, conditioned on measured growth rate, with model-in-the-loop parameter identification to separate biopart-associated properties from host-dependent effects. Using structured combinatorial libraries, we identify mechanistically interpretable, transferable parameters for plasmid origins, promoters and ribosome binding sites. In particular, we define an intrinsic translation initiation capacity that captures the dominant RBS-associated contribution to translation while context-dependent expression emerges from host physiology and local sequence context. The resulting parameterisation accurately predicts protein synthesis across physiological conditions, supports incremental library expansion, and reveals localised failures of modularity, providing a scalable foundation for predictive host-aware design in synthetic biology. Model-based design in synthetic biology is limited because bioparts are typically characterised by relative metrics that vary across genetic and physiological contexts. Here, using combinatorial libraries of plasmid-based expression constructs and a host-aware E. coli digital twin, the authors identify intrinsic biopart parameters that provide a basis for more predictive model-based design of synthetic gene circuits.