I've been dealing with a seasonal cold recently, and while thinking about what my white blood cells are doing inside my body right now, I started thinking about cellular automata.
That somehow led me to this idea.
I came across this short video showing a white blood cell surrounding and engulfing bacteria, and it made me wonder:
Could we create a cellular automaton that produces something resembling this behavior?
The basic idea would be to create a very simple artificial immune system with a few basic entities:
- White blood cells
- Pathogens / bacteria
- Empty space
The white blood cell would ideally be able to:
- Detect a nearby pathogen
- Move toward it
- Surround it
- Engulf it
- Eventually eliminate or absorb it
But I think the really interesting part is that none of this should be explicitly programmed as a sequence of actions.
Instead, the behavior could emerge from
simple local rules.
For example, rather than telling a white blood cell "go toward the bacterium", we could define rules based only on its local neighborhood. If a pathogen is nearby, certain cells at the boundary could change state or extend toward it. Once the pathogen becomes partially surrounded, further local interactions could cause the surrounding structure to continue growing around it.
Eventually, complete engulfment could emerge naturally from those local interactions.
Simple local rules → complex, biological-looking behavior.
I think it would be particularly interesting if the white blood cell were not just a predefined pattern moving around, but a genuinely
shape-shifting object.
Its "body" could dynamically extend pseudopod-like structures, change shape, surround pathogens, and eventually absorb them, all as an emergent consequence of the rules.
And perhaps the idea could be taken further.
For example:
- Pathogens could reproduce: 1 → 2 → 4 → 8...
- White blood cells could have a limited capacity for engulfment.
- Movement or engulfment could have some kind of "cost".
- A white blood cell could become temporarily inactive after engulfing a pathogen.
- Additional white blood cells could appear or be recruited when the pathogen population becomes large.
At that point, it could become a very simple
artificial immune system, with something like:
Pathogen reproduction vs. immune-cell response
But the main question is much simpler:
Can a small set of local rules produce behavior that looks surprisingly similar to phagocytosis?
I haven't worked out the rules yet. I'm mainly throwing the idea out there because I thought it might make an interesting Life-like experiment.
Has anyone seen a cellular automaton that produces similar behavior, or perhaps something that could be adapted to this idea?
I'd be particularly interested in seeing how far we could get using
only local interactions, without explicitly programming the overall behavior.

- PhagocytosisWhiteBloodCellsEngagedInWork.gif (7.4 MiB) Viewed 18 times