The purpose of this research programme is to contribute to the solution of a major problem in today’s systems biology: namely, the difficulty to bring mechanistic modelling to bear on full scale systems. Statistical, and experimental techniques have scaled up considerably in the last two decades. Mechanistic modelling, on the other hand, is still confined in much smaller scales. Witness the painstakingly slow scaling up of cellular signalling models, despite their central role in cell response. To address the problem, we build on a new modelling methodology, called the rule-based approach (RB), pioneered by the PI of this proposal, and hailed (in a recent Nature Methods article) as the “harbinger of an entirely new way of representing and studying cellular networks”. By exploiting the modularity of biological agents, RB breaks through the combinatorial challenge of describing and simulating signalling systems. But with the possibility of writing and running larger models, new questions come to the fore. To bring mechanistic modelling to the next level requires: innovative knowledge representation techniques to anchor modelling in the data-side of systems bi- ology; new means to tame the complexity of, and reason about, the parameter space of models; new concepts to identify meaningful observables in the highly stochastic behaviour of large and combinatorial models; and clean and structured languages to comprehend spatial aspects of the biological phenomenology. The realism accrued by working at larger scales gets one closer to the bottom-up reconstruction of behaviours at the heart of systems biology, and to an understanding of the computational architecture of complex biological networks. This research programme, firmly grounded in the mathematics of programming language semantics and formal methods, extends the RB approach so as to address all of the above needs, and deliver an integrated modelling framework where full scale mechanistic modelling is achievable.