Synthetic circuit evolution
“The logic of biological regulatory systems abides not by Hegelian
laws but, like the workings of computers, by the propositional alge-
bra of George Boole.” (Francois Jacob & Jacques Monod)
I have just finished reading this article, Bacterial Computing, written by Martyn Amos of Manchester Metropolitan University. The idea that we might harness the rules of nature to perform logical computations in living systems has fascinated me for a couple of years now but I’ve never properly come across any description of it taking place. I was pretty stoked when I found this article.
To give you a gist, Martyn describes how bacteria can be used for such boolean algebra. They can be made to do this by exploiting the many signaling mechanisms that are used in genetic regulatory networks. That is, if you draw a flow diagram of how expression of one gene manipulates the expression of another via signaling and protein building then, for certain genes, signaling mechanisms and proteins, you end up drawing a circuit, much like an electronic one. The behaviour of this circuit can be very complex and Martyn doesn’t fail to exhibit some fascinating examples of how these circuits can be designed to perform distinct functions such as producing an oscillating fluorescence, taking photographs and targeting tumor cells.
A problem is that when you insert these genetic networks into living cells so that they can operate, the results, that is, the ability of the cells to perform a desired task, are disappointingly unreliable due to the inherent noisyness of cellular processes. Martyn describes how some scientists get around this problem by deliberately mutating the cells. In most cases, the mutation gives even worse results but, occasionally, the mutation increases the performance of the cells. By keeping those cells that perform better and discarding those that perform worse, the scientists create an artificial evolutionary process over short time scales by which the cells are selected (in a survival of the fittest manner) to perform the desired task reliably. How cool is that!
So I was thinking - can this be achieved not just in electrical circuits, not just in genetic regulatory networks but in social, behavioural networks as well? Given a large number of behaving organisms (so as to smooth out noise) which can communicate between each other (say, via pheromones) and that will react to certain environmental pressures, you could indeed create a logical behavioural circuit. One could actually harness the predictability of collective behaviour (for example, swarming in locusts) such that, given a change in the environment, any predictable change in behaviour would report, to the experimenter, that very change in the environment with a certain degree of reliability.