A new agent-based platform for simulating multicellular biocircuits with conjugative plasmids

Resumen

Synthetic biology is a recently born discipline, finding its origins in the year 2000 at the MIT Artificial Intelligence lab. It emerges as a discipline in which engineering is applied to biological components and systems. As per its nature, synthetic biology is an interdisciplinary field in which engineers and biologists work together to design and implement circuits that either seek to recreate life or display an unnatural behavior for a specific purpose. Until recently, this engineering of biology has been carried out in an informal manner, as bioCAD and simulation tools were not available to aid in the design of biocircuits. A trial-and-error cycle along with the biologist intuition are the guidelines for a long process in reaching the desired biocircuit behavior. This places synthetic biology at a stage that can be compared to microelectronics in the 1980s. Trial- and-error is a slow method, simulators and bioCAD tools are a requirement for speeding up the process. The design-build-test-learn cycle is the approach most synthetic biologists currently use to undertake the task. This cycle takes full advantage of the software tools to speed up the process. In the design phase of this cycle, the scientist works on a design and elaborates on its functionality in a theoretical manner. It is at this stage that computer simulations become useful, as they reproduce the basic dynamics of the designs. These programs serve the purpose of assisting in identifying flaws and ruling out incorrect designs. They also help in directing and estimating parameters to implement designs in the wet lab. One goal is to minimize the costs in time and money due to experimentation. In this scenario, there is a clear need for fast, accurate and simple to use simulators. Agent based Models (AbMs), also known as Individual based Models (IbMs), are a type of models implemented by simulators that focus on the behavior of each individual in a colony (in this context, it is a colony of cells). These models are alternatives to Ordinary Differential Equation (ODE) or stochastic (Gillespie) based models. AbMs provide a better representation of how bacterial populations interact and behave in a spatial manner, whilst ODE and Gillespie models offer a more accurate simulation of dynamics at an intracellular level. Bacterial colonies grow in a Petri dish in a 2D manner. Each bacterium can be seen as an individual agent that communicates and interacts with other bacteria through environmental chemical signals or by exchanging genetic messages (plasmids). Each cell follows probabilistic rules that direct its growth, reproduction, communication, death, or any action it executes. This accounts for the whole bacterium behavior. These probabilistic rules yield the emergence of a global population-level behavior. AbM simulators are well-suited for assessing and predicting multicellular synthetic circuit behavior. This leads to three research questions on the matter: 1) What potential do current AbM simulators have for simulating novel multicellular genetic circuits using bacterial conjugation? 2) How should the identified shortcomings be overcome to enable the simulation of novel multicellular genetic circuits using bacterial conjugation? 3) What novel designs relating to spatial and/or temporal patterns can the newly enhanced platform prototype? When this work began, no AbM simulator was capable of simulating multicellular circuits in the PLASWIRES project. gro was selected as the AbM framework to be improved and extended. Following this choice, and the scaling complexity in simulations, a roadmap of features to complement it was established: • Improving the simulator speed: this was done by implementing CellEngine, a new physics engine featuring a shoving algorithm tailored for bacterial colonies. • Adding bacterial conjugation as an intercell communication mechanism: implemented directly into the gro source code, being a key requirement for simulating PLASWIRES circuits. • Simplifying the gene expression process in the simulator and simulation specification: CellPro is a new module that simplifies gene expression through binary protein dynamics. ProSpec defines a new scalable specification language for describing complex genetic circuits. • Adding nutrient uptake and modulating cell growth: a Monod based nutrient uptake simulator module, CellNutrient, was developed to couple with gro. • Improving the environmental signal capabilities: CellSignals packages a new and extended version of environmental signal simulation for gro. gro was improved substiantially upon inclusion of the proposed features. It is now capable of simulating 105 bacteria in a matter of minutes (instead of a week). This newly-acquired speed of the simulator places it as a very fast prototyping tool for a wide range of systems and synthetic biology designs. Additionally, gro now provides built-in genetic specification and dynamics, nutrient consumption, improved environmental signal capabilities and bacterial conjugation. This platform, along with the mentioned modules are open-source and can be found at: https://github.com/liaupm/GRO-LIA. Realism of the simulations produced by the platform has increased as high cell counts are now attainable and can make use of the new features. Outcomes of these simulations accurately shed light on the behavior of multicellular circuits, fulfilling their purpose in prototyping for early assessment of circuit design. This work also presents novel multicellular designs as an application of the tool for implementation and analysis of multicellular synthetic circuit designs. They mainly refer to spatial and/or temporal patterns, and make full use of the newly included features of gro. Future developments on the simulator include automating simulation specification, offering a web-based version of the simulator, connecting to existing standards such as SBOL or SBML and adding onto the existing features by implementing bacteriophage infection or gene editing capabilities such as a CRISPR/Cas9 system. gro has become a tool that is useful for a broader audience than its previous version. Multicellular systems biology, and, in particular, microbial evolutionary biology and microbiota programming are examples of other fields that could benefit from a new platform like the one presented in the current document.