Some of the ideas of this project came from our attempts to build a modelling and simulation environment for understanding the behaviour of adult stem cells. One existing model viewed the stem cells (in a niche such as in the bone marrow) as a CA system. If you look at the simulation you will see that a constant there are always differentiated cells being produced, and that the stem cell population never diminishes greatly (it is self re-newing). These are two of the fundamental capabilities of a system of stem cells.

Cellular automata version of stem cells

The problem with the CA version is that the empty space (the black holes) have to do some computational work (they have counters and know what kinds of cells are next to them). This would seem to be biologically counter intuitive. Moreover, division is not supported. We have remodelled this simulation using agents, who have perceptions, and who can divide into empty spaces. In order to get exactly the same simulation we require 5-way division! This clearly suggests problems with the original model.

Multi-agent version of stem cells with 5-way division

However, if we allow 2-way division only. We can get a simlar model that can be demonstrated to have the same properties. (in terms of producing determined cells and maintaing the stem cell population).

Multi-agent version allowing only 2-way division

The idea of modelling a stem cell as an agent is an attractive one. Most importantly it enables a biologically intuitive model of stem cells to be developed which can then help dialogue between experimental biologists and people from the formal modelling/simulation community. We have found a surprising lack of inter-disciplinary work where experimentalists are working with modellers and this is partly due to no shared language for communication. You can have a look at some papers on the web site that explore some of the features of interdisciplinary research.

 Constructing Voronoi diagram of UK cities in CA model of reaction-diffusion system (see Adamatzky, Ben De Lacy Costello, Asai, “Reaction-diffusion computers”, Elsevier, 2005, for details)  

 Constructing Voronoi diagram of UK cities in lattice swarm model of reaction-diffusion system (see Adamatzky, Computing in Nonlinear Media, IoP Publishing, 2001, for details) GIF ANIMATION