Multicellular reservoir computing is a simulation I’ve done of an arbitrary populations of cells in order to analyze their computational capacity. In other words, can an arbitrarily given population of cells be used as a computational device with its enviroment as input.

The project concludes that indeed this is possible and analyzes factors that maximize computational performance. And what kind of useful computation can you do in this scenario? Given that the environment (meaning, the surrounding molecules that the cells can register) is the input, you can perhaps analyze the content of the environment or, more ambitiously, engineer the cells to manipulate it.

The exciting part about the project was that we made minimal assumptions about the interal gene “wiring” of the cells (how the genes are regulated), since that was randomly generated in every simulation per cell strain. And despite that, we were able to derive useful computation from the cell population by simply observing the environment and their secreted signals.