Agent-based model

Proliferation of bacteriophage infection in an Escherichia coli culture is a complicated phenomenon. Even more complex is the behavior of our synthetic Escherichia coli population which fights against the infection.

In order to test if the intracellular defence system was effectively contending versus phages at population level, we implement an [http://en.wikipedia.org/wiki/Agent-based_model agent-based simulation] using the program [http://ccl.northwestern.edu/netlogo/ NetLogo].
Agent-based models have the advantage of being easily constructed. This approach allows us to recreate complex processes and the interactions of thousands of objects (for instance, bacteria and phages) in parallel, thus exploring their effects in the community (the culture) as a whole.

In an agent-based model each object has its own variables and states:

AGENTS

• Bacteria
• Phages

VARIABLES:

Bacteria's variables

• Period
• Latency period
• Time to release AHL
• Time to die by system

Phages' variables

• Burst size
• Bacteriophage decay rate
• Effective infection probability (Bacteriophage Adsorption Rate)

RULES

The objects interact with each other through a series of defined rules:

• Bacteria' movement
• Phages' movement
• Infection process
• Suicide system activation
• Diffusion of AHL
• Antisense production when AHL reach to the activation threshold

ASSUMPTIONS

For simplicity the model make certain assumptions:

• The time is discrete
• Only one phage infect one bacteria
• Phages moves at random
• Antisense avoid totally the phage production

TO RUN THE MODEL

This model consists of two types of experiments:

• NEGATIVE CONTROL.

Simulate a T7 infection of wild-type Escherichia coli. When a phage encounters a bacterium, it attaches itself to the cell wall of the bacterium and start an infection. The phage takes over the E.coli’s metabolic machinery in order to ensemble multiple copies of itself. Within a time tau (latency period) the host bacteria die, releasing phage particles ready to infect nearby cells. Phages win the game.

• FIGHT FIRE WITH FIRE.

When an E. coli is infected by a phage, the warning device and the suicide system are activated. The burst size of the phage is severely diminished (according to our BSD prediction). At the same time, the infected bacteria produces GFP (that’s why they are green :p) and simultaneously release AHL to the medium. AHL act as a signal to warn the neighbor bacteria of the presence of phages in the vicinity. Advised bacteria begins to synthesize antisense RNA molecules that protect it against phages. Warned bacteria additionally produces RFP indicating that they are protected.

Agents' states

E. coli that has not yet been infected.	E. coli wild-type that has been infected by the phage.
Synthetic E. coli that has been infected by the bacteriophage. BSD predict a diminished number of phage produced by this infected cell. Additionally, GFP is produced.	Synthetic E. coli advised of the existence of an infection in the neighborhood. RNA antisense is produced, preventing the reproduction of phage if the bacteria become infected. Additionally, RFP is produced.

How to use it?

Click the SETUP button to setup the world. Then click on the experiment that you deserve to run (NEGATIVE CONTOL or FIGHT FIRE WITH FIRE). You can control the number of bacteria and the number of initial infections with the slide bars. Also, with the switch display-AHL, you can choose to observe or hide the AHL diffusion.

E. coli that has not yet been infected.

E. coli wt that has been infected by the phage.

Synthetic E. coli that has been infected by the phage.

Synthetic E. coli which has already been advised of the existence of an infection in the neighborhood.

AGENTS’ STATES

E. coli that has not yet been infected.

E. coli wt that has been infected by the phage.

Synthetic E. coli that has been infected by the phage.

Bacterium which has already been advised of the existence of an infection in the neighborhood.

Important to note: In this model, for clarity, we have neglected bacterial duplication since we are only interested in the local dynamic of the population. In this model we have neglected bacterial duplication because we are only interested in the local dynamic of the population. If duplication is incorporated we lose to some extent the