Model-based Design of Synthetic Biological Circuits

Abstract

The rational design of synthetic gene circuits has led to many successful applications over the past two decades. Computational methods enable the design of synthetic biological circuits demonstrating a specific dynamic behavior. Current methods rest on the assembly of parts characterized in different contexts, which often fail to operate as predicted when combined, or when they operate in a different context. Increasingly complex constructs revealed that analogies to electronics design such as modularity and ‘plug-and-play’ composition are of limited use: biology is less well characterized, more context-dependent, and overall less predictable. In this thesis, we first summarize the main conceptual challenges of synthetic circuit design to highlight recent progress towards more tailored, context-aware computational design methods for synthetic biology. We then follow three approaches to take up the challenge of designing and constructing functional synthetic circuits despite context-dependencies. First, we contribute to the design of beta-cell mimetic designer cells for diabetes treatment, with a model that puts the synthetic circuit in the context of its host cell and host organism, filling the gap between in vitro and in vivo experiments. Second, we introduce a circuit design method that compensates for parts’ uncertainty by identifying circuit topologies whose behavior is robust to variations in parameters. We validate this method in silico with the well-known case study of three-node circuits able to achieve biochemical adaptation. Third, taking advantage of the scalability of our computational method, we extend it to create TopoDesign, the most important contribution of this thesis: a novel rapid prototyping approach that finds the circuit topology that is the most likely to work with already available parts in the context of a specific lab. We validate TopoDesign experimentally by applying it to the case of designing and constructing a synthetic circuit with a novel, non-intuitive behavior: a decoder that discriminates between short and long pulses of an intracellular signal. By responding to short pulses of signals only, and leaving aside no inputs or long inputs, such a decoder would enable us to multiplex endogenous pathways for synthetic biology purposes, in a way that does not hinder the natural function of the pathway. Alternatively, the decoder could respond appropriately to, for example, physiological signals on different time-scales in future biomedical applications. TopoDesign allowed us to explore more than 4’000 possible circuit architectures, design targeted experiments, and then rationally select a single circuit for implementation. We constructed the selected circuit and it demonstrated the desired short-pulse decoding behavior. While most of the current model-based design methods focus on the design of logical circuits with common well-characterized parts, our approach expands the scope of reliable design to more complex dynamic behaviors. Its two main features, Bayesian robustness design, and rapid prototyping, account respectively for parts' biological variability, and uncertain knowledge about the parts available in different contexts.