"

From 2010.igem.org

Revision as of 16:02, 16 September 2010 (view source) Ttodd (Talk | contribs) m ← Older edit		Revision as of 13:57, 13 October 2010 (view source) Ttodd (Talk | contribs) Newer edit →
Line 13:		Line 13:
	==Modelling Strategy==		==Modelling Strategy==
		+	Modelling has an important role in synthetic biology. Being able to predict how a novel gene
		+	regulatory network (GRN) will behave is crucial. Before one commits the time and resources to
		+	actually constructing such a network out of BioBricks (or other pieces of genetic material) it is
		+	important to know if it is stable, if the time-scales that it operates on are relevant and if it will
		+	produce the responses that one is looking for. Many approaches exist to model GRNs, the most
		+	popular being coupled ordinary differential equations (ODEs), stochastic networks and graphical
		+	models.
		+
		+	For the levels of expression considered in this project the coupled ODE approach is the most
		+	valid, it assumes that quantities of repressors, promoters, proteins etc. are high enough to be rep-
		+	resented as being well mixed and continuous. This aspect of modelling is already well implemented
		+	in our BSim modelling software, the aspect that we improved upon in this years release is the
		+	geometric aspect of the modelling.

---

Revision as of 13:57, 13 October 2010

Contents 1 Subpage Structure 2 Modelling Strategy 2.1 Overall Aims 2.2 Modelling aims 3 Current activity 4 Summary 4.1 Modelling Capsules 4.2 BSim GUI

Subpage Structure

Team:BCCS-Bristol/Modelling/GRN

Team:BCCS-Bristol/Modelling/BSIM

Team:BCCS-Bristol/Modelling/Data_Analysis

Modelling Strategy

Modelling has an important role in synthetic biology. Being able to predict how a novel gene regulatory network (GRN) will behave is crucial. Before one commits the time and resources to actually constructing such a network out of BioBricks (or other pieces of genetic material) it is important to know if it is stable, if the time-scales that it operates on are relevant and if it will produce the responses that one is looking for. Many approaches exist to model GRNs, the most popular being coupled ordinary differential equations (ODEs), stochastic networks and graphical models.

For the levels of expression considered in this project the coupled ODE approach is the most valid, it assumes that quantities of repressors, promoters, proteins etc. are high enough to be rep- resented as being well mixed and continuous. This aspect of modelling is already well implemented in our BSim modelling software, the aspect that we improved upon in this years release is the geometric aspect of the modelling.

Overall Aims

Our modelling strategy needs to tie in with our larger plan for this year’s project. We are working on a soil analysis system for use in agriculture, based on an E. coli chassis. Our bacteria will need to be able to:

• Reliably detect nitrate levels in soil
• Signal using GFP

Additionally, our bacteria need to be:

• Detectable
• Contained

We need to be able to detect the signal from our bacteria using equipment that can feasibly used on a farm. Containing our bacteria means they cannot interfere with the outside world in undesirable ways.

Modelling aims

The primary function of our modelling team is to use computational simulations to assist and inform the work of our biologists. Currently, we plan to achieve this by working on these areas:

• GRN models
  • Simple GRN model (Nitrate Detector → GFP)
  • Composite GRN (Detect several nutrients → Respond with a combination of signals)
• Behaviour of Capsules compared to direct application to soil (spray)
  • Nitrate Diffusion into capsules and cells
    • Modify BSim to allow for irregular boundaries: Currently, BSim divides space into discrete cubic blocks. This makes defining curved or spherical boundaries difficult. We are working on an algorithm to approximate curved surfaces in a cubic environment.
  • Analyse data from microscope experiments
  • Detectability of Capsules
    • Quorum Sensing to amplify signal: The outermost bacteria in a capsule may be able to detect nitrates much more effectively than those buried deep within the capsule. If the outer bacteria signalled that they detect nitrates, the inner bacteria could detect this and express GFP. This would effectively amplify GFP expression in the capsule.
    • Multiple populations & sensitivity: By using populations of different sensitivities, we can use the total GFP expression as a more accurate measure of the concentration of nitrates in the soil.
    • Constant expression of RFP, detect ratio of GFP to RFP: If our bacteria express one colour constantly, and express another colour in response to nitrate detection, the ratio of the two colours can be used to quantify the level of nitrate detection.

Current activity

We have been working on these areas in the last few weeks:

• Implementing boundaries and chemical fields in BSim
• Diffusion through capsules
• Fitting parameters to GRN
• Soil analysis
• GUI
• Collaborations

Summary

The modelling team should have two main outputs. These are:

• Modelling Capsules
• Developing BSim GUI and encouraging collaboration

Modelling Capsules

Containing our bacteria within a semi-permeable capsule is one of the more unique aspects of our project. We can use our models to inform the design of these capsules; the size of the capsule and thickness of its coating are dependent on our estimates of diffusion through them, and the detectability of the GFP within them.

BSim GUI

BSim is currently an excellent modelling tool, but not widely used. By giving it a Graphical User Interface (GUI), we can make it more accessible to the wider synthetic biology community, who may lack JAVA knowledge. To help test and improve our GUI, we will make it available to several other IGEM teams. This will also constitute a modelling collaboration, which is a sufficient criterion for a gold medal.