Team:Edinburgh/Modeling

Motivation

Synthetic biologists often design genetic circuitry in isolation, taking little consideration of the host cells in which these circuits will operate. They tend to create specific, local models which don't capture the circuits' interactions with other host components. This is an oversimplification because the circuit genes and products interact with the host cell in various ways:

• The circuit is dependent upon the resources and machinery available to the cell – so if resources are scarce, this is likely to hinder the circuit transcription and translation.
• The cell needs to replicate, translate and transcribe the additional genes inserted into it and this draws upon the host’s resources which could otherwise be used for metabolism and growth. As a result, if the circuit is long or the genes on it are overexpressed, this can slow down the growth of the host cell.
• The gene products of the circuit might interact with the cell metabolism in an undesirable manner. For example, they might be toxic to the host. Alternatively, some of the host's metabolic enzymes might inhibit the circuit's production rate; an obviously unwanted side effect.

Failing to take account of those interactions and their consequences at the design stage can cause designs to fail or be sub-optimal.

Goals

With this in mind, we decided to introduce the concept of whole-cell modeling to iGEM: modeling the entire cell and capturing key factors of its life cycle and metabolism. A very abstract, high-level cell "template" could be made thus, or instead a very detailed, richly-informative model, depending on the data available and on the specific application. We can then insert specific circuit models into this whole-cell model and see how the circuit would operate in the context of the cell. In this way, we can create better-informed designs, which have a symbiotic rather than a parasitic relationship with their host.

It would be even better to have a living breathing computer cell that is accessible to everyone, despite its turbid programmatic depths. The way to achieve this would be to have a universal simulation platform with a modular nature, in which different modules can be easily added and removed. We can have the whole cell model running on this platform, and we can easily add any specific circuit models as black-box modules.

This idea can be extended to make the whole-cell model itself into a module which can run in the simulation platform. Thus it would be possible to choose a whole cell from a library of models, or to program your own one. It may also be possible to turn on/off some features of the whole cell model, customizing it to be more coarse-grained or more fine-grained according to your preferences.

With this ambitious view in mind, we set off to explore this idea and make a first few steps towards implementing it.

Goals

• Pick or program a suitable simulation platform
• Implement a whole cell model on it
• Insert a specific circuit model into it – modeling the waste water treatment system that the biology team is making
• Make predictions about the functioning of the circuit in the context of the cell